#!/usr/bin/env perl
-#******************************************************************************#
-#* Copyright(c) 2012, Intel Corp. *#
-#* Developers and authors: *#
-#* Shay Gueron (1, 2), and Vlad Krasnov (1) *#
-#* (1) Intel Architecture Group, Microprocessor and Chipset Development, *#
-#* Israel Development Center, Haifa, Israel *#
-#* (2) University of Haifa *#
-#******************************************************************************#
-#* This submission to OpenSSL is to be made available under the OpenSSL *#
-#* license, and only to the OpenSSL project, in order to allow integration *#
-#* into the publicly distributed code. ? *#
-#* The use of this code, or portions of this code, or concepts embedded in *#
-#* this code, or modification of this code and/or algorithm(s) in it, or the *#
-#* use of this code for any other purpose than stated above, requires special *#
-#* licensing. *#
-#******************************************************************************#
-#******************************************************************************#
-#* DISCLAIMER: *#
-#* THIS SOFTWARE IS PROVIDED BY THE CONTRIBUTORS AND THE COPYRIGHT OWNERS *#
-#* ``AS IS''. ANY EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED *#
-#* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR *#
-#* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE CONTRIBUTORS OR THE COPYRIGHT*#
-#* OWNERS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, *#
-#* OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF *#
-#* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS *#
-#* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN *#
-#* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) *#
-#* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE *#
-#* POSSIBILITY OF SUCH DAMAGE. *#
-#******************************************************************************#
-#* Reference: *#
-#* [1] S. Gueron, "Efficient Software Implementations of Modular *#
-#* Exponentiation", http://eprint.iacr.org/2011/239 *#
-#* [2] S. Gueron, V. Krasnov. "Speeding up Big-Numbers Squaring". *#
-#* IEEE Proceedings of 9th International Conference on Information *#
-#* Technology: New Generations (ITNG 2012), 821-823 (2012). *#
-#* [3] S. Gueron, Efficient Software Implementations of Modular Exponentiation*#
-#* Journal of Cryptographic Engineering 2:31-43 (2012). *#
-#* [4] S. Gueron, V. Krasnov: "[PATCH] Efficient and side channel analysis *#
-#* resistant 512-bit and 1024-bit modular exponentiation for optimizing *#
-#* RSA1024 and RSA2048 on x86_64 platforms", *#
-#* http://rt.openssl.org/Ticket/Display.html?id=2582&user=guest&pass=guest*#
-################################################################################
+##############################################################################
+# #
+# Copyright (c) 2012, Intel Corporation #
+# #
+# All rights reserved. #
+# #
+# Redistribution and use in source and binary forms, with or without #
+# modification, are permitted provided that the following conditions are #
+# met: #
+# #
+# * Redistributions of source code must retain the above copyright #
+# notice, this list of conditions and the following disclaimer. #
+# #
+# * Redistributions in binary form must reproduce the above copyright #
+# notice, this list of conditions and the following disclaimer in the #
+# documentation and/or other materials provided with the #
+# distribution. #
+# #
+# * Neither the name of the Intel Corporation nor the names of its #
+# contributors may be used to endorse or promote products derived from #
+# this software without specific prior written permission. #
+# #
+# #
+# THIS SOFTWARE IS PROVIDED BY INTEL CORPORATION ""AS IS"" AND ANY #
+# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE #
+# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR #
+# PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL INTEL CORPORATION OR #
+# CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, #
+# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, #
+# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR #
+# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF #
+# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING #
+# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS #
+# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. #
+# #
+##############################################################################
+# Developers and authors: #
+# Shay Gueron (1, 2), and Vlad Krasnov (1) #
+# (1) Intel Architecture Group, Microprocessor and Chipset Development, #
+# Israel Development Center, Haifa, Israel #
+# (2) University of Haifa #
+##############################################################################
+# Reference: #
+# [1] S. Gueron, "Efficient Software Implementations of Modular #
+# Exponentiation", http://eprint.iacr.org/2011/239 #
+# [2] S. Gueron, V. Krasnov. "Speeding up Big-Numbers Squaring". #
+# IEEE Proceedings of 9th International Conference on Information #
+# Technology: New Generations (ITNG 2012), 821-823 (2012). #
+# [3] S. Gueron, Efficient Software Implementations of Modular Exponentiation#
+# Journal of Cryptographic Engineering 2:31-43 (2012). #
+# [4] S. Gueron, V. Krasnov: "[PATCH] Efficient and side channel analysis #
+# resistant 512-bit and 1024-bit modular exponentiation for optimizing #
+# RSA1024 and RSA2048 on x86_64 platforms", #
+# http://rt.openssl.org/Ticket/Display.html?id=2582&user=guest&pass=guest#
+##############################################################################
# While original submission covers 512- and 1024-bit exponentiation,
# this module is limited to 512-bit version only (and as such
#
# (*) rsax engine and fips numbers are presented for reference
# purposes;
-# (**) you might notice MULX code below, strangely enough gain is
-# marginal, which is why code remains disabled;
+# (**) MULX was attempted, but found to give only marginal improvement;
$flavour = shift;
$output = shift;
( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
die "can't locate x86_64-xlate.pl";
-open OUT,"| $^X $xlate $flavour $output";
+open OUT,"| \"$^X\" $xlate $flavour $output";
*STDOUT=*OUT;
+if (`$ENV{CC} -Wa,-v -c -o /dev/null -x assembler /dev/null 2>&1`
+ =~ /GNU assembler version ([2-9]\.[0-9]+)/) {
+ $addx = ($1>=2.23);
+}
+
+if (!$addx && $win64 && ($flavour =~ /nasm/ || $ENV{ASM} =~ /nasm/) &&
+ `nasm -v 2>&1` =~ /NASM version ([2-9]\.[0-9]+)/) {
+ $addx = ($1>=2.10);
+}
+
+if (!$addx && $win64 && ($flavour =~ /masm/ || $ENV{ASM} =~ /ml64/) &&
+ `ml64 2>&1` =~ /Version ([0-9]+)\./) {
+ $addx = ($1>=12);
+}
+
+if (!$addx && `$ENV{CC} -v 2>&1` =~ /((?:^clang|LLVM) version|.*based on LLVM) ([3-9])\.([0-9]+)/) {
+ my $ver = $2 + $3/100.0; # 3.1->3.01, 3.10->3.10
+ $addx = ($ver>=3.03);
+}
+
($out, $inp, $mod) = ("%rdi", "%rsi", "%rbp"); # common internal API
{
my ($out,$inp,$mod,$n0,$times) = ("%rdi","%rsi","%rdx","%rcx","%r8d");
$code.=<<___;
.text
+.extern OPENSSL_ia32cap_P
+
.globl rsaz_512_sqr
-.type rsaz_512_sqr,\@function,4
+.type rsaz_512_sqr,\@function,5
.align 32
rsaz_512_sqr: # 25-29% faster than rsaz_512_mul
push %rbx
movq ($inp), %rdx
movq 8($inp), %rax
movq $n0, 128(%rsp)
+___
+$code.=<<___ if ($addx);
+ movl \$0x80100,%r11d
+ andl OPENSSL_ia32cap_P+8(%rip),%r11d
+ cmpl \$0x80100,%r11d # check for MULX and ADO/CX
+ je .Loop_sqrx
+___
+$code.=<<___;
jmp .Loop_sqr
.align 32
.Loop_sqr:
movl $times,128+8(%rsp)
-___
-if (1) {
-$code.=<<___;
#first iteration
movq %rdx, %rbx
mulq %rdx
movq %r13, 112(%rsp)
movq %r14, 120(%rsp)
+
+ movq (%rsp), %r8
+ movq 8(%rsp), %r9
+ movq 16(%rsp), %r10
+ movq 24(%rsp), %r11
+ movq 32(%rsp), %r12
+ movq 40(%rsp), %r13
+ movq 48(%rsp), %r14
+ movq 56(%rsp), %r15
+
+ call __rsaz_512_reduce
+
+ addq 64(%rsp), %r8
+ adcq 72(%rsp), %r9
+ adcq 80(%rsp), %r10
+ adcq 88(%rsp), %r11
+ adcq 96(%rsp), %r12
+ adcq 104(%rsp), %r13
+ adcq 112(%rsp), %r14
+ adcq 120(%rsp), %r15
+ sbbq %rcx, %rcx
+
+ call __rsaz_512_subtract
+
+ movq %r8, %rdx
+ movq %r9, %rax
+ movl 128+8(%rsp), $times
+ movq $out, $inp
+
+ decl $times
+ jnz .Loop_sqr
___
-} else {
+if ($addx) {
$code.=<<___;
+ jmp .Lsqr_tail
+
+.align 32
+.Loop_sqrx:
+ movl $times,128+8(%rsp)
movq $out, %xmm0 # off-load
+ movq %rbp, %xmm1 # off-load
#first iteration
mulx %rax, %r8, %r9
mulx 16($inp), %rcx, %r10
+ xor %rbp, %rbp # cf=0, of=0
mulx 24($inp), %rax, %r11
- add %rcx, %r9
+ adcx %rcx, %r9
mulx 32($inp), %rcx, %r12
- adc %rax, %r10
+ adcx %rax, %r10
mulx 40($inp), %rax, %r13
- adc %rcx, %r11
+ adcx %rcx, %r11
- mulx 48($inp), %rcx, %r14
- adc %rax, %r12
+ .byte 0xc4,0x62,0xf3,0xf6,0xb6,0x30,0x00,0x00,0x00 # mulx 48($inp), %rcx, %r14
+ adcx %rax, %r12
+ adcx %rcx, %r13
- mulx 56($inp), %rax, %r15
- adc %rcx, %r13
- mov %r9, %rcx
- adc %rax, %r14
- adc \$0, %r15
+ .byte 0xc4,0x62,0xfb,0xf6,0xbe,0x38,0x00,0x00,0x00 # mulx 56($inp), %rax, %r15
+ adcx %rax, %r14
+ adcx %rbp, %r15 # %rbp is 0
+ mov %r9, %rcx
shld \$1, %r8, %r9
shl \$1, %r8
+ xor %ebp, %ebp
mulx %rdx, %rax, %rdx
- add %rdx, %r8
- adc \$0, %r9
+ adcx %rdx, %r8
+ mov 8($inp), %rdx
+ adcx %rbp, %r9
mov %rax, (%rsp)
mov %r8, 8(%rsp)
#second iteration
- mov 8($inp), %rdx
mulx 16($inp), %rax, %rbx
+ adox %rax, %r10
+ adcx %rbx, %r11
- mulx 24($inp), $out, %r8
- add %rax, %r10
- adc %rbx, %r11
- adc \$0, %r8
+ .byte 0xc4,0x62,0xc3,0xf6,0x86,0x18,0x00,0x00,0x00 # mulx 24($inp), $out, %r8
+ adox $out, %r11
+ adcx %r8, %r12
mulx 32($inp), %rax, %rbx
- add $out, %r11
- adc %r8, %r12
- adc \$0, %rbx
+ adox %rax, %r12
+ adcx %rbx, %r13
mulx 40($inp), $out, %r8
- add %rax, %r12
- adc %rbx, %r13
- adc \$0, %r8
+ adox $out, %r13
+ adcx %r8, %r14
- mulx 48($inp), %rax, %rbx
- add $out, %r13
- adc %r8, %r14
- adc \$0, %rbx
+ .byte 0xc4,0xe2,0xfb,0xf6,0x9e,0x30,0x00,0x00,0x00 # mulx 48($inp), %rax, %rbx
+ adox %rax, %r14
+ adcx %rbx, %r15
- mulx 56($inp), $out, %r8
- add %rax, %r14
- adc %rbx, %r15
- mov %r11, %rbx
- adc \$0, %r8
- add $out, %r15
- adc \$0, %r8
+ .byte 0xc4,0x62,0xc3,0xf6,0x86,0x38,0x00,0x00,0x00 # mulx 56($inp), $out, %r8
+ adox $out, %r15
+ adcx %rbp, %r8
+ adox %rbp, %r8
+ mov %r11, %rbx
shld \$1, %r10, %r11
shld \$1, %rcx, %r10
+ xor %ebp,%ebp
mulx %rdx, %rax, %rcx
- add %rax, %r9
- adc %rcx, %r10
- adc \$0, %r11
+ mov 16($inp), %rdx
+ adcx %rax, %r9
+ adcx %rcx, %r10
+ adcx %rbp, %r11
mov %r9, 16(%rsp)
- mov %r10, 24(%rsp)
+ .byte 0x4c,0x89,0x94,0x24,0x18,0x00,0x00,0x00 # mov %r10, 24(%rsp)
#third iteration
- mov 16($inp), %rdx
- mulx 24($inp), $out, %r9
+ .byte 0xc4,0x62,0xc3,0xf6,0x8e,0x18,0x00,0x00,0x00 # mulx 24($inp), $out, %r9
+ adox $out, %r12
+ adcx %r9, %r13
mulx 32($inp), %rax, %rcx
- add $out, %r12
- adc %r9, %r13
- adc \$0, %rcx
+ adox %rax, %r13
+ adcx %rcx, %r14
mulx 40($inp), $out, %r9
- add %rax, %r13
- adc %rcx, %r14
- adc \$0, %r9
+ adox $out, %r14
+ adcx %r9, %r15
- mulx 48($inp), %rax, %rcx
- add $out, %r14
- adc %r9, %r15
- adc \$0, %rcx
+ .byte 0xc4,0xe2,0xfb,0xf6,0x8e,0x30,0x00,0x00,0x00 # mulx 48($inp), %rax, %rcx
+ adox %rax, %r15
+ adcx %rcx, %r8
- mulx 56($inp), $out, %r9
- add %rax, %r15
- adc %rcx, %r8
- mov %r13, %rcx
- adc \$0, %r9
- add $out, %r8
- adc \$0, %r9
+ .byte 0xc4,0x62,0xc3,0xf6,0x8e,0x38,0x00,0x00,0x00 # mulx 56($inp), $out, %r9
+ adox $out, %r8
+ adcx %rbp, %r9
+ adox %rbp, %r9
+ mov %r13, %rcx
shld \$1, %r12, %r13
shld \$1, %rbx, %r12
+ xor %ebp, %ebp
mulx %rdx, %rax, %rdx
- add %rax, %r11
- adc %rdx, %r12
- adc \$0, %r13
+ adcx %rax, %r11
+ adcx %rdx, %r12
+ mov 24($inp), %rdx
+ adcx %rbp, %r13
mov %r11, 32(%rsp)
- mov %r12, 40(%rsp)
+ .byte 0x4c,0x89,0xa4,0x24,0x28,0x00,0x00,0x00 # mov %r12, 40(%rsp)
#fourth iteration
- mov 24($inp), %rdx
- mulx 32($inp), %rax, %rbx
+ .byte 0xc4,0xe2,0xfb,0xf6,0x9e,0x20,0x00,0x00,0x00 # mulx 32($inp), %rax, %rbx
+ adox %rax, %r14
+ adcx %rbx, %r15
mulx 40($inp), $out, %r10
- add %rax, %r14
- adc %rbx, %r15
- adc \$0, %r10
+ adox $out, %r15
+ adcx %r10, %r8
mulx 48($inp), %rax, %rbx
- add $out, %r15
- adc %r10, %r8
- adc \$0, %rbx
+ adox %rax, %r8
+ adcx %rbx, %r9
mulx 56($inp), $out, %r10
- add %rax, %r8
- adc \$0, %rbx
- add $out, %r9
- adc \$0, %r10
- add %rbx, %r9
- mov %r15, %rbx
- adc \$0, %r10
+ adox $out, %r9
+ adcx %rbp, %r10
+ adox %rbp, %r10
+ .byte 0x66
+ mov %r15, %rbx
shld \$1, %r14, %r15
shld \$1, %rcx, %r14
+ xor %ebp, %ebp
mulx %rdx, %rax, %rdx
- add %rax, %r13
- adc %rdx, %r14
- adc \$0, %r15
+ adcx %rax, %r13
+ adcx %rdx, %r14
+ mov 32($inp), %rdx
+ adcx %rbp, %r15
mov %r13, 48(%rsp)
mov %r14, 56(%rsp)
#fifth iteration
- mov 32($inp), %rdx
- mulx 40($inp), $out, %r11
+ .byte 0xc4,0x62,0xc3,0xf6,0x9e,0x28,0x00,0x00,0x00 # mulx 40($inp), $out, %r11
+ adox $out, %r8
+ adcx %r11, %r9
mulx 48($inp), %rax, %rcx
- add $out, %r8
- adc %r11, %r9
- adc \$0, %rcx
+ adox %rax, %r9
+ adcx %rcx, %r10
mulx 56($inp), $out, %r11
- add %rax, %r9
- adc %rcx, %r10
- adc \$0, %r11
- add $out, %r10
- adc \$0, %r11
+ adox $out, %r10
+ adcx %rbp, %r11
+ adox %rbp, %r11
mov %r9, %rcx
shld \$1, %r8, %r9
shld \$1, %rbx, %r8
+ xor %ebp, %ebp
mulx %rdx, %rax, %rdx
- add %rax, %r15
- adc %rdx, %r8
- adc \$0, %r9
+ adcx %rax, %r15
+ adcx %rdx, %r8
+ mov 40($inp), %rdx
+ adcx %rbp, %r9
mov %r15, 64(%rsp)
mov %r8, 72(%rsp)
#sixth iteration
- mov 40($inp), %rdx
- mulx 48($inp), %rax, %rbx
+ .byte 0xc4,0xe2,0xfb,0xf6,0x9e,0x30,0x00,0x00,0x00 # mulx 48($inp), %rax, %rbx
+ adox %rax, %r10
+ adcx %rbx, %r11
- mulx 56($inp), $out, %r12
- add %rax, %r10
- adc %rbx, %r11
- adc \$0, %r12
- add $out, %r11
- adc \$0, %r12
+ .byte 0xc4,0x62,0xc3,0xf6,0xa6,0x38,0x00,0x00,0x00 # mulx 56($inp), $out, %r12
+ adox $out, %r11
+ adcx %rbp, %r12
+ adox %rbp, %r12
mov %r11, %rbx
shld \$1, %r10, %r11
shld \$1, %rcx, %r10
+ xor %ebp, %ebp
mulx %rdx, %rax, %rdx
- add %rax, %r9
- adc %rdx, %r10
- adc \$0, %r11
+ adcx %rax, %r9
+ adcx %rdx, %r10
+ mov 48($inp), %rdx
+ adcx %rbp, %r11
mov %r9, 80(%rsp)
mov %r10, 88(%rsp)
#seventh iteration
- mov 48($inp), %rdx
- mulx 56($inp), %rax, %r13
- add %rax, %r12
- adc \$0, %r13
+ .byte 0xc4,0x62,0xfb,0xf6,0xae,0x38,0x00,0x00,0x00 # mulx 56($inp), %rax, %r13
+ adox %rax, %r12
+ adox %rbp, %r13
xor %r14, %r14
shld \$1, %r13, %r14
shld \$1, %r12, %r13
shld \$1, %rbx, %r12
+ xor %ebp, %ebp
mulx %rdx, %rax, %rdx
- add %rax, %r11
- adc %rdx, %r12
- adc \$0, %r13
+ adcx %rax, %r11
+ adcx %rdx, %r12
+ mov 56($inp), %rdx
+ adcx %rbp, %r13
- mov %r11, 96(%rsp)
- mov %r12, 104(%rsp)
+ .byte 0x4c,0x89,0x9c,0x24,0x60,0x00,0x00,0x00 # mov %r11, 96(%rsp)
+ .byte 0x4c,0x89,0xa4,0x24,0x68,0x00,0x00,0x00 # mov %r12, 104(%rsp)
#eighth iteration
- mov 56($inp), %rdx
mulx %rdx, %rax, %rdx
- add %rax, %r13
- adc \$0, %rdx
-
+ adox %rax, %r13
+ adox %rbp, %rdx
+
+ .byte 0x66
add %rdx, %r14
movq %r13, 112(%rsp)
movq %r14, 120(%rsp)
movq %xmm0, $out
-___
-}
-$code.=<<___;
+ movq %xmm1, %rbp
+
+ movq 128(%rsp), %rdx # pull $n0
movq (%rsp), %r8
movq 8(%rsp), %r9
movq 16(%rsp), %r10
movq 48(%rsp), %r14
movq 56(%rsp), %r15
- call _rsaz_512_reduce
+ call __rsaz_512_reducex
addq 64(%rsp), %r8
adcq 72(%rsp), %r9
adcq 120(%rsp), %r15
sbbq %rcx, %rcx
- call _rsaz_512_subtract
+ call __rsaz_512_subtract
movq %r8, %rdx
movq %r9, %rax
movq $out, $inp
decl $times
- jnz .Loop_sqr
+ jnz .Loop_sqrx
+
+.Lsqr_tail:
+___
+}
+$code.=<<___;
leaq 128+24+48(%rsp), %rax
movq -48(%rax), %r15
movq $out, %xmm0 # off-load arguments
movq $mod, %xmm1
movq $n0, 128(%rsp)
-
+___
+$code.=<<___ if ($addx);
+ movl \$0x80100,%r11d
+ andl OPENSSL_ia32cap_P+8(%rip),%r11d
+ cmpl \$0x80100,%r11d # check for MULX and ADO/CX
+ je .Lmulx
+___
+$code.=<<___;
+ movq ($bp), %rbx # pass b[0]
movq $bp, %rbp # pass argument
call __rsaz_512_mul
movq 48(%rsp), %r14
movq 56(%rsp), %r15
- call _rsaz_512_reduce
+ call __rsaz_512_reduce
+___
+$code.=<<___ if ($addx);
+ jmp .Lmul_tail
+
+.align 32
+.Lmulx:
+ movq $bp, %rbp # pass argument
+ movq ($bp), %rdx # pass b[0]
+ call __rsaz_512_mulx
+
+ movq %xmm0, $out
+ movq %xmm1, %rbp
+ movq 128(%rsp), %rdx # pull $n0
+ movq (%rsp), %r8
+ movq 8(%rsp), %r9
+ movq 16(%rsp), %r10
+ movq 24(%rsp), %r11
+ movq 32(%rsp), %r12
+ movq 40(%rsp), %r13
+ movq 48(%rsp), %r14
+ movq 56(%rsp), %r15
+
+ call __rsaz_512_reducex
+.Lmul_tail:
+___
+$code.=<<___;
addq 64(%rsp), %r8
adcq 72(%rsp), %r9
adcq 80(%rsp), %r10
adcq 120(%rsp), %r15
sbbq %rcx, %rcx
- call _rsaz_512_subtract
+ call __rsaz_512_subtract
leaq 128+24+48(%rsp), %rax
movq -48(%rax), %r15
push %r14
push %r15
- subq \$128+24, %rsp
+ subq \$`128+24+($win64?0xb0:0)`, %rsp
+___
+$code.=<<___ if ($win64);
+ movaps %xmm6,0xa0(%rsp)
+ movaps %xmm7,0xb0(%rsp)
+ movaps %xmm8,0xc0(%rsp)
+ movaps %xmm9,0xd0(%rsp)
+ movaps %xmm10,0xe0(%rsp)
+ movaps %xmm11,0xf0(%rsp)
+ movaps %xmm12,0x100(%rsp)
+ movaps %xmm13,0x110(%rsp)
+ movaps %xmm14,0x120(%rsp)
+ movaps %xmm15,0x130(%rsp)
+___
+$code.=<<___;
.Lmul_gather4_body:
- movl 64($bp,$pwr,4), %eax
- movq $out, %xmm0 # off-load arguments
- movl ($bp,$pwr,4), %ebx
- movq $mod, %xmm1
- movq $n0, 128(%rsp)
+ movd $pwr,%xmm8
+ movdqa .Linc+16(%rip),%xmm1 # 00000002000000020000000200000002
+ movdqa .Linc(%rip),%xmm0 # 00000001000000010000000000000000
+
+ pshufd \$0,%xmm8,%xmm8 # broadcast $power
+ movdqa %xmm1,%xmm7
+ movdqa %xmm1,%xmm2
+___
+########################################################################
+# calculate mask by comparing 0..15 to $power
+#
+for($i=0;$i<4;$i++) {
+$code.=<<___;
+ paddd %xmm`$i`,%xmm`$i+1`
+ pcmpeqd %xmm8,%xmm`$i`
+ movdqa %xmm7,%xmm`$i+3`
+___
+}
+for(;$i<7;$i++) {
+$code.=<<___;
+ paddd %xmm`$i`,%xmm`$i+1`
+ pcmpeqd %xmm8,%xmm`$i`
+___
+}
+$code.=<<___;
+ pcmpeqd %xmm8,%xmm7
+
+ movdqa 16*0($bp),%xmm8
+ movdqa 16*1($bp),%xmm9
+ movdqa 16*2($bp),%xmm10
+ movdqa 16*3($bp),%xmm11
+ pand %xmm0,%xmm8
+ movdqa 16*4($bp),%xmm12
+ pand %xmm1,%xmm9
+ movdqa 16*5($bp),%xmm13
+ pand %xmm2,%xmm10
+ movdqa 16*6($bp),%xmm14
+ pand %xmm3,%xmm11
+ movdqa 16*7($bp),%xmm15
+ leaq 128($bp), %rbp
+ pand %xmm4,%xmm12
+ pand %xmm5,%xmm13
+ pand %xmm6,%xmm14
+ pand %xmm7,%xmm15
+ por %xmm10,%xmm8
+ por %xmm11,%xmm9
+ por %xmm12,%xmm8
+ por %xmm13,%xmm9
+ por %xmm14,%xmm8
+ por %xmm15,%xmm9
+
+ por %xmm9,%xmm8
+ pshufd \$0x4e,%xmm8,%xmm9
+ por %xmm9,%xmm8
+___
+$code.=<<___ if ($addx);
+ movl \$0x80100,%r11d
+ andl OPENSSL_ia32cap_P+8(%rip),%r11d
+ cmpl \$0x80100,%r11d # check for MULX and ADO/CX
+ je .Lmulx_gather
+___
+$code.=<<___;
+ movq %xmm8,%rbx
+
+ movq $n0, 128(%rsp) # off-load arguments
+ movq $out, 128+8(%rsp)
+ movq $mod, 128+16(%rsp)
- shlq \$32, %rax
- or %rax, %rbx
movq ($ap), %rax
movq 8($ap), %rcx
- leaq 128($bp,$pwr,4), %rbp
mulq %rbx # 0 iteration
movq %rax, (%rsp)
movq %rcx, %rax
movq %rdx, %r8
mulq %rbx
- movd (%rbp), %xmm4
addq %rax, %r8
movq 16($ap), %rax
movq %rdx, %r9
adcq \$0, %r9
mulq %rbx
- movd 64(%rbp), %xmm5
addq %rax, %r9
movq 24($ap), %rax
movq %rdx, %r10
adcq \$0, %r10
mulq %rbx
- pslldq \$4, %xmm5
addq %rax, %r10
movq 32($ap), %rax
movq %rdx, %r11
adcq \$0, %r11
mulq %rbx
- por %xmm5, %xmm4
addq %rax, %r11
movq 40($ap), %rax
movq %rdx, %r12
adcq \$0, %r13
mulq %rbx
- leaq 128(%rbp), %rbp
addq %rax, %r13
movq 56($ap), %rax
movq %rdx, %r14
adcq \$0, %r14
mulq %rbx
- movq %xmm4, %rbx
addq %rax, %r14
movq ($ap), %rax
movq %rdx, %r15
.align 32
.Loop_mul_gather:
+ movdqa 16*0(%rbp),%xmm8
+ movdqa 16*1(%rbp),%xmm9
+ movdqa 16*2(%rbp),%xmm10
+ movdqa 16*3(%rbp),%xmm11
+ pand %xmm0,%xmm8
+ movdqa 16*4(%rbp),%xmm12
+ pand %xmm1,%xmm9
+ movdqa 16*5(%rbp),%xmm13
+ pand %xmm2,%xmm10
+ movdqa 16*6(%rbp),%xmm14
+ pand %xmm3,%xmm11
+ movdqa 16*7(%rbp),%xmm15
+ leaq 128(%rbp), %rbp
+ pand %xmm4,%xmm12
+ pand %xmm5,%xmm13
+ pand %xmm6,%xmm14
+ pand %xmm7,%xmm15
+ por %xmm10,%xmm8
+ por %xmm11,%xmm9
+ por %xmm12,%xmm8
+ por %xmm13,%xmm9
+ por %xmm14,%xmm8
+ por %xmm15,%xmm9
+
+ por %xmm9,%xmm8
+ pshufd \$0x4e,%xmm8,%xmm9
+ por %xmm9,%xmm8
+ movq %xmm8,%rbx
+
mulq %rbx
addq %rax, %r8
movq 8($ap), %rax
adcq \$0, %r8
mulq %rbx
- movd (%rbp), %xmm4
addq %rax, %r9
movq 16($ap), %rax
adcq \$0, %rdx
adcq \$0, %r9
mulq %rbx
- movd 64(%rbp), %xmm5
addq %rax, %r10
movq 24($ap), %rax
adcq \$0, %rdx
adcq \$0, %r10
mulq %rbx
- pslldq \$4, %xmm5
addq %rax, %r11
movq 32($ap), %rax
adcq \$0, %rdx
adcq \$0, %r11
mulq %rbx
- por %xmm5, %xmm4
addq %rax, %r12
movq 40($ap), %rax
adcq \$0, %rdx
adcq \$0, %r14
mulq %rbx
- movq %xmm4, %rbx
addq %rax, %r15
movq ($ap), %rax
adcq \$0, %rdx
movq %rdx, %r15
adcq \$0, %r15
- leaq 128(%rbp), %rbp
leaq 8(%rdi), %rdi
decl %ecx
movq %r14, 48(%rdi)
movq %r15, 56(%rdi)
- movq %xmm0, $out
- movq %xmm1, %rbp
+ movq 128+8(%rsp), $out
+ movq 128+16(%rsp), %rbp
movq (%rsp), %r8
movq 8(%rsp), %r9
movq 48(%rsp), %r14
movq 56(%rsp), %r15
- call _rsaz_512_reduce
+ call __rsaz_512_reduce
+___
+$code.=<<___ if ($addx);
+ jmp .Lmul_gather_tail
+
+.align 32
+.Lmulx_gather:
+ movq %xmm8,%rdx
+
+ mov $n0, 128(%rsp) # off-load arguments
+ mov $out, 128+8(%rsp)
+ mov $mod, 128+16(%rsp)
+
+ mulx ($ap), %rbx, %r8 # 0 iteration
+ mov %rbx, (%rsp)
+ xor %edi, %edi # cf=0, of=0
+
+ mulx 8($ap), %rax, %r9
+
+ mulx 16($ap), %rbx, %r10
+ adcx %rax, %r8
+
+ mulx 24($ap), %rax, %r11
+ adcx %rbx, %r9
+ mulx 32($ap), %rbx, %r12
+ adcx %rax, %r10
+
+ mulx 40($ap), %rax, %r13
+ adcx %rbx, %r11
+
+ mulx 48($ap), %rbx, %r14
+ adcx %rax, %r12
+
+ mulx 56($ap), %rax, %r15
+ adcx %rbx, %r13
+ adcx %rax, %r14
+ .byte 0x67
+ mov %r8, %rbx
+ adcx %rdi, %r15 # %rdi is 0
+
+ mov \$-7, %rcx
+ jmp .Loop_mulx_gather
+
+.align 32
+.Loop_mulx_gather:
+ movdqa 16*0(%rbp),%xmm8
+ movdqa 16*1(%rbp),%xmm9
+ movdqa 16*2(%rbp),%xmm10
+ movdqa 16*3(%rbp),%xmm11
+ pand %xmm0,%xmm8
+ movdqa 16*4(%rbp),%xmm12
+ pand %xmm1,%xmm9
+ movdqa 16*5(%rbp),%xmm13
+ pand %xmm2,%xmm10
+ movdqa 16*6(%rbp),%xmm14
+ pand %xmm3,%xmm11
+ movdqa 16*7(%rbp),%xmm15
+ leaq 128(%rbp), %rbp
+ pand %xmm4,%xmm12
+ pand %xmm5,%xmm13
+ pand %xmm6,%xmm14
+ pand %xmm7,%xmm15
+ por %xmm10,%xmm8
+ por %xmm11,%xmm9
+ por %xmm12,%xmm8
+ por %xmm13,%xmm9
+ por %xmm14,%xmm8
+ por %xmm15,%xmm9
+
+ por %xmm9,%xmm8
+ pshufd \$0x4e,%xmm8,%xmm9
+ por %xmm9,%xmm8
+ movq %xmm8,%rdx
+
+ .byte 0xc4,0x62,0xfb,0xf6,0x86,0x00,0x00,0x00,0x00 # mulx ($ap), %rax, %r8
+ adcx %rax, %rbx
+ adox %r9, %r8
+
+ mulx 8($ap), %rax, %r9
+ adcx %rax, %r8
+ adox %r10, %r9
+
+ mulx 16($ap), %rax, %r10
+ adcx %rax, %r9
+ adox %r11, %r10
+
+ .byte 0xc4,0x62,0xfb,0xf6,0x9e,0x18,0x00,0x00,0x00 # mulx 24($ap), %rax, %r11
+ adcx %rax, %r10
+ adox %r12, %r11
+
+ mulx 32($ap), %rax, %r12
+ adcx %rax, %r11
+ adox %r13, %r12
+
+ mulx 40($ap), %rax, %r13
+ adcx %rax, %r12
+ adox %r14, %r13
+
+ .byte 0xc4,0x62,0xfb,0xf6,0xb6,0x30,0x00,0x00,0x00 # mulx 48($ap), %rax, %r14
+ adcx %rax, %r13
+ .byte 0x67
+ adox %r15, %r14
+
+ mulx 56($ap), %rax, %r15
+ mov %rbx, 64(%rsp,%rcx,8)
+ adcx %rax, %r14
+ adox %rdi, %r15
+ mov %r8, %rbx
+ adcx %rdi, %r15 # cf=0
+
+ inc %rcx # of=0
+ jnz .Loop_mulx_gather
+
+ mov %r8, 64(%rsp)
+ mov %r9, 64+8(%rsp)
+ mov %r10, 64+16(%rsp)
+ mov %r11, 64+24(%rsp)
+ mov %r12, 64+32(%rsp)
+ mov %r13, 64+40(%rsp)
+ mov %r14, 64+48(%rsp)
+ mov %r15, 64+56(%rsp)
+
+ mov 128(%rsp), %rdx # pull arguments
+ mov 128+8(%rsp), $out
+ mov 128+16(%rsp), %rbp
+
+ mov (%rsp), %r8
+ mov 8(%rsp), %r9
+ mov 16(%rsp), %r10
+ mov 24(%rsp), %r11
+ mov 32(%rsp), %r12
+ mov 40(%rsp), %r13
+ mov 48(%rsp), %r14
+ mov 56(%rsp), %r15
+
+ call __rsaz_512_reducex
+
+.Lmul_gather_tail:
+___
+$code.=<<___;
addq 64(%rsp), %r8
adcq 72(%rsp), %r9
adcq 80(%rsp), %r10
adcq 120(%rsp), %r15
sbbq %rcx, %rcx
- call _rsaz_512_subtract
+ call __rsaz_512_subtract
leaq 128+24+48(%rsp), %rax
+___
+$code.=<<___ if ($win64);
+ movaps 0xa0-0xc8(%rax),%xmm6
+ movaps 0xb0-0xc8(%rax),%xmm7
+ movaps 0xc0-0xc8(%rax),%xmm8
+ movaps 0xd0-0xc8(%rax),%xmm9
+ movaps 0xe0-0xc8(%rax),%xmm10
+ movaps 0xf0-0xc8(%rax),%xmm11
+ movaps 0x100-0xc8(%rax),%xmm12
+ movaps 0x110-0xc8(%rax),%xmm13
+ movaps 0x120-0xc8(%rax),%xmm14
+ movaps 0x130-0xc8(%rax),%xmm15
+ lea 0xb0(%rax),%rax
+___
+$code.=<<___;
movq -48(%rax), %r15
movq -40(%rax), %r14
movq -32(%rax), %r13
push %r14
push %r15
+ mov $pwr, $pwr
subq \$128+24, %rsp
.Lmul_scatter4_body:
- leaq ($tbl,$pwr,4), $tbl
+ leaq ($tbl,$pwr,8), $tbl
movq $out, %xmm0 # off-load arguments
movq $mod, %xmm1
movq $tbl, %xmm2
movq $n0, 128(%rsp)
movq $out, %rbp
+___
+$code.=<<___ if ($addx);
+ movl \$0x80100,%r11d
+ andl OPENSSL_ia32cap_P+8(%rip),%r11d
+ cmpl \$0x80100,%r11d # check for MULX and ADO/CX
+ je .Lmulx_scatter
+___
+$code.=<<___;
+ movq ($out),%rbx # pass b[0]
call __rsaz_512_mul
movq %xmm0, $out
movq 48(%rsp), %r14
movq 56(%rsp), %r15
- call _rsaz_512_reduce
+ call __rsaz_512_reduce
+___
+$code.=<<___ if ($addx);
+ jmp .Lmul_scatter_tail
+
+.align 32
+.Lmulx_scatter:
+ movq ($out), %rdx # pass b[0]
+ call __rsaz_512_mulx
+
+ movq %xmm0, $out
+ movq %xmm1, %rbp
+
+ movq 128(%rsp), %rdx # pull $n0
+ movq (%rsp), %r8
+ movq 8(%rsp), %r9
+ movq 16(%rsp), %r10
+ movq 24(%rsp), %r11
+ movq 32(%rsp), %r12
+ movq 40(%rsp), %r13
+ movq 48(%rsp), %r14
+ movq 56(%rsp), %r15
+
+ call __rsaz_512_reducex
+.Lmul_scatter_tail:
+___
+$code.=<<___;
addq 64(%rsp), %r8
adcq 72(%rsp), %r9
adcq 80(%rsp), %r10
movq %xmm2, $inp
sbbq %rcx, %rcx
- call _rsaz_512_subtract
-
- movl %r8d, 64*0($inp) # scatter
- shrq \$32, %r8
- movl %r9d, 64*2($inp)
- shrq \$32, %r9
- movl %r10d, 64*4($inp)
- shrq \$32, %r10
- movl %r11d, 64*6($inp)
- shrq \$32, %r11
- movl %r12d, 64*8($inp)
- shrq \$32, %r12
- movl %r13d, 64*10($inp)
- shrq \$32, %r13
- movl %r14d, 64*12($inp)
- shrq \$32, %r14
- movl %r15d, 64*14($inp)
- shrq \$32, %r15
- movl %r8d, 64*1($inp)
- movl %r9d, 64*3($inp)
- movl %r10d, 64*5($inp)
- movl %r11d, 64*7($inp)
- movl %r12d, 64*9($inp)
- movl %r13d, 64*11($inp)
- movl %r14d, 64*13($inp)
- movl %r15d, 64*15($inp)
+ call __rsaz_512_subtract
+
+ movq %r8, 128*0($inp) # scatter
+ movq %r9, 128*1($inp)
+ movq %r10, 128*2($inp)
+ movq %r11, 128*3($inp)
+ movq %r12, 128*4($inp)
+ movq %r13, 128*5($inp)
+ movq %r14, 128*6($inp)
+ movq %r15, 128*7($inp)
leaq 128+24+48(%rsp), %rax
movq -48(%rax), %r15
subq \$128+24, %rsp
.Lmul_by_one_body:
+___
+$code.=<<___ if ($addx);
+ movl OPENSSL_ia32cap_P+8(%rip),%eax
+___
+$code.=<<___;
movq $mod, %rbp # reassign argument
movq $n0, 128(%rsp)
movdqa %xmm0, 64(%rsp)
movdqa %xmm0, 80(%rsp)
movdqa %xmm0, 96(%rsp)
-
- call _rsaz_512_reduce
-
+___
+$code.=<<___ if ($addx);
+ andl \$0x80100,%eax
+ cmpl \$0x80100,%eax # check for MULX and ADO/CX
+ je .Lby_one_callx
+___
+$code.=<<___;
+ call __rsaz_512_reduce
+___
+$code.=<<___ if ($addx);
+ jmp .Lby_one_tail
+.align 32
+.Lby_one_callx:
+ movq 128(%rsp), %rdx # pull $n0
+ call __rsaz_512_reducex
+.Lby_one_tail:
+___
+$code.=<<___;
movq %r8, ($out)
movq %r9, 8($out)
movq %r10, 16($out)
.size rsaz_512_mul_by_one,.-rsaz_512_mul_by_one
___
}
-{ # _rsaz_512_reduce
+{ # __rsaz_512_reduce
#
# input: %r8-%r15, %rbp - mod, 128(%rsp) - n0
# output: %r8-%r15
# clobbers: everything except %rbp and %rdi
$code.=<<___;
-.type _rsaz_512_reduce,\@abi-omnipotent
+.type __rsaz_512_reduce,\@abi-omnipotent
.align 32
-_rsaz_512_reduce:
-___
-if (1) {
-$code.=<<___;
+__rsaz_512_reduce:
movq %r8, %rbx
imulq 128+8(%rsp), %rbx
movq 0(%rbp), %rax
adcq \$0, %rdx
addq %r11, %r10
movq 128+8(%rsp), %rsi
+ #movq %rdx, %r11
+ #adcq \$0, %r11
+ adcq \$0, %rdx
movq %rdx, %r11
- adcq \$0, %r11
mulq %rbx
addq %rax, %r12
decl %ecx
jne .Lreduction_loop
+
+ ret
+.size __rsaz_512_reduce,.-__rsaz_512_reduce
___
-} else {
+}
+if ($addx) {
+ # __rsaz_512_reducex
+ #
+ # input: %r8-%r15, %rbp - mod, 128(%rsp) - n0
+ # output: %r8-%r15
+ # clobbers: everything except %rbp and %rdi
$code.=<<___;
- movq 128+8(%rsp), %rdx # pull $n0
+.type __rsaz_512_reducex,\@abi-omnipotent
+.align 32
+__rsaz_512_reducex:
+ #movq 128+8(%rsp), %rdx # pull $n0
imulq %r8, %rdx
+ xorq %rsi, %rsi # cf=0,of=0
movl \$8, %ecx
- jmp .Lreduction_loop
+ jmp .Lreduction_loopx
.align 32
-.Lreduction_loop:
- neg %r8
+.Lreduction_loopx:
+ mov %r8, %rbx
mulx 0(%rbp), %rax, %r8
- adc %r9, %r8
+ adcx %rbx, %rax
+ adox %r9, %r8
mulx 8(%rbp), %rax, %r9
- adc \$0, %r9
- add %rax, %r8
- adc %r10, %r9
-
- mulx 16(%rbp), %rax, %r10
- adc \$0, %r10
- mov 128+8(%rsp), %rbx # pull $n0
- imul %r8, %rbx
- add %rax, %r9
- adc %r11, %r10
-
- mulx 24(%rbp), %rax, %r11
- adc \$0, %r11
- add %rax, %r10
- adc %r12, %r11
-
- mulx 32(%rbp), %rax, %r12
- adc \$0, %r12
- add %rax, %r11
- adc %r13, %r12
+ adcx %rax, %r8
+ adox %r10, %r9
+
+ mulx 16(%rbp), %rbx, %r10
+ adcx %rbx, %r9
+ adox %r11, %r10
+
+ mulx 24(%rbp), %rbx, %r11
+ adcx %rbx, %r10
+ adox %r12, %r11
+
+ .byte 0xc4,0x62,0xe3,0xf6,0xa5,0x20,0x00,0x00,0x00 # mulx 32(%rbp), %rbx, %r12
+ mov %rdx, %rax
+ mov %r8, %rdx
+ adcx %rbx, %r11
+ adox %r13, %r12
+
+ mulx 128+8(%rsp), %rbx, %rdx
+ mov %rax, %rdx
mulx 40(%rbp), %rax, %r13
- adc \$0, %r13
- add %rax, %r12
- adc %r14, %r13
+ adcx %rax, %r12
+ adox %r14, %r13
- mulx 48(%rbp), %rax, %r14
- adc \$0, %r14
- add %rax, %r13
- adc %r15, %r14
+ .byte 0xc4,0x62,0xfb,0xf6,0xb5,0x30,0x00,0x00,0x00 # mulx 48(%rbp), %rax, %r14
+ adcx %rax, %r13
+ adox %r15, %r14
mulx 56(%rbp), %rax, %r15
mov %rbx, %rdx
- adc \$0, %r15
- add %rax, %r14
- adc \$0, %r15
+ adcx %rax, %r14
+ adox %rsi, %r15 # %rsi is 0
+ adcx %rsi, %r15 # cf=0
+
+ decl %ecx # of=0
+ jne .Lreduction_loopx
- dec %ecx
- jne .Lreduction_loop
-___
-}
-$code.=<<___;
ret
-.size _rsaz_512_reduce,.-_rsaz_512_reduce
+.size __rsaz_512_reducex,.-__rsaz_512_reducex
___
}
-{ # _rsaz_512_subtract
+{ # __rsaz_512_subtract
# input: %r8-%r15, %rdi - $out, %rbp - $mod, %rcx - mask
# output:
# clobbers: everything but %rdi, %rsi and %rbp
$code.=<<___;
-.type _rsaz_512_subtract,\@abi-omnipotent
+.type __rsaz_512_subtract,\@abi-omnipotent
.align 32
-_rsaz_512_subtract:
+__rsaz_512_subtract:
movq %r8, ($out)
movq %r9, 8($out)
movq %r10, 16($out)
movq %r15, 56($out)
ret
-.size _rsaz_512_subtract,.-_rsaz_512_subtract
+.size __rsaz_512_subtract,.-__rsaz_512_subtract
___
}
{ # __rsaz_512_mul
#
# input: %rsi - ap, %rbp - bp
- # ouput:
+ # output:
# clobbers: everything
my ($ap,$bp) = ("%rsi","%rbp");
$code.=<<___;
__rsaz_512_mul:
leaq 8(%rsp), %rdi
- movq ($bp), %rbx
movq ($ap), %rax
mulq %rbx
movq %rax, (%rdi)
.size __rsaz_512_mul,.-__rsaz_512_mul
___
}
+if ($addx) {
+ # __rsaz_512_mulx
+ #
+ # input: %rsi - ap, %rbp - bp
+ # output:
+ # clobbers: everything
+my ($ap,$bp,$zero) = ("%rsi","%rbp","%rdi");
+$code.=<<___;
+.type __rsaz_512_mulx,\@abi-omnipotent
+.align 32
+__rsaz_512_mulx:
+ mulx ($ap), %rbx, %r8 # initial %rdx preloaded by caller
+ mov \$-6, %rcx
+
+ mulx 8($ap), %rax, %r9
+ movq %rbx, 8(%rsp)
+
+ mulx 16($ap), %rbx, %r10
+ adc %rax, %r8
+
+ mulx 24($ap), %rax, %r11
+ adc %rbx, %r9
+
+ mulx 32($ap), %rbx, %r12
+ adc %rax, %r10
+
+ mulx 40($ap), %rax, %r13
+ adc %rbx, %r11
+
+ mulx 48($ap), %rbx, %r14
+ adc %rax, %r12
+
+ mulx 56($ap), %rax, %r15
+ mov 8($bp), %rdx
+ adc %rbx, %r13
+ adc %rax, %r14
+ adc \$0, %r15
+
+ xor $zero, $zero # cf=0,of=0
+ jmp .Loop_mulx
+
+.align 32
+.Loop_mulx:
+ movq %r8, %rbx
+ mulx ($ap), %rax, %r8
+ adcx %rax, %rbx
+ adox %r9, %r8
+
+ mulx 8($ap), %rax, %r9
+ adcx %rax, %r8
+ adox %r10, %r9
+
+ mulx 16($ap), %rax, %r10
+ adcx %rax, %r9
+ adox %r11, %r10
+
+ mulx 24($ap), %rax, %r11
+ adcx %rax, %r10
+ adox %r12, %r11
+
+ .byte 0x3e,0xc4,0x62,0xfb,0xf6,0xa6,0x20,0x00,0x00,0x00 # mulx 32($ap), %rax, %r12
+ adcx %rax, %r11
+ adox %r13, %r12
+
+ mulx 40($ap), %rax, %r13
+ adcx %rax, %r12
+ adox %r14, %r13
+
+ mulx 48($ap), %rax, %r14
+ adcx %rax, %r13
+ adox %r15, %r14
+
+ mulx 56($ap), %rax, %r15
+ movq 64($bp,%rcx,8), %rdx
+ movq %rbx, 8+64-8(%rsp,%rcx,8)
+ adcx %rax, %r14
+ adox $zero, %r15
+ adcx $zero, %r15 # cf=0
+
+ inc %rcx # of=0
+ jnz .Loop_mulx
+
+ movq %r8, %rbx
+ mulx ($ap), %rax, %r8
+ adcx %rax, %rbx
+ adox %r9, %r8
+
+ .byte 0xc4,0x62,0xfb,0xf6,0x8e,0x08,0x00,0x00,0x00 # mulx 8($ap), %rax, %r9
+ adcx %rax, %r8
+ adox %r10, %r9
+
+ .byte 0xc4,0x62,0xfb,0xf6,0x96,0x10,0x00,0x00,0x00 # mulx 16($ap), %rax, %r10
+ adcx %rax, %r9
+ adox %r11, %r10
+
+ mulx 24($ap), %rax, %r11
+ adcx %rax, %r10
+ adox %r12, %r11
+
+ mulx 32($ap), %rax, %r12
+ adcx %rax, %r11
+ adox %r13, %r12
+
+ mulx 40($ap), %rax, %r13
+ adcx %rax, %r12
+ adox %r14, %r13
+
+ .byte 0xc4,0x62,0xfb,0xf6,0xb6,0x30,0x00,0x00,0x00 # mulx 48($ap), %rax, %r14
+ adcx %rax, %r13
+ adox %r15, %r14
+
+ .byte 0xc4,0x62,0xfb,0xf6,0xbe,0x38,0x00,0x00,0x00 # mulx 56($ap), %rax, %r15
+ adcx %rax, %r14
+ adox $zero, %r15
+ adcx $zero, %r15
+
+ mov %rbx, 8+64-8(%rsp)
+ mov %r8, 8+64(%rsp)
+ mov %r9, 8+64+8(%rsp)
+ mov %r10, 8+64+16(%rsp)
+ mov %r11, 8+64+24(%rsp)
+ mov %r12, 8+64+32(%rsp)
+ mov %r13, 8+64+40(%rsp)
+ mov %r14, 8+64+48(%rsp)
+ mov %r15, 8+64+56(%rsp)
+
+ ret
+.size __rsaz_512_mulx,.-__rsaz_512_mulx
+___
+}
{
my ($out,$inp,$power)= $win64 ? ("%rcx","%rdx","%r8d") : ("%rdi","%rsi","%edx");
$code.=<<___;
.type rsaz_512_scatter4,\@abi-omnipotent
.align 16
rsaz_512_scatter4:
- leaq ($out,$power,4), $out
+ leaq ($out,$power,8), $out
movl \$8, %r9d
jmp .Loop_scatter
.align 16
.Loop_scatter:
movq ($inp), %rax
leaq 8($inp), $inp
- movl %eax, ($out)
- shrq \$32, %rax
- movl %eax, 64($out)
+ movq %rax, ($out)
leaq 128($out), $out
decl %r9d
jnz .Loop_scatter
.type rsaz_512_gather4,\@abi-omnipotent
.align 16
rsaz_512_gather4:
- leaq ($inp,$power,4), $inp
+___
+$code.=<<___ if ($win64);
+.LSEH_begin_rsaz_512_gather4:
+ .byte 0x48,0x81,0xec,0xa8,0x00,0x00,0x00 # sub $0xa8,%rsp
+ .byte 0x0f,0x29,0x34,0x24 # movaps %xmm6,(%rsp)
+ .byte 0x0f,0x29,0x7c,0x24,0x10 # movaps %xmm7,0x10(%rsp)
+ .byte 0x44,0x0f,0x29,0x44,0x24,0x20 # movaps %xmm8,0x20(%rsp)
+ .byte 0x44,0x0f,0x29,0x4c,0x24,0x30 # movaps %xmm9,0x30(%rsp)
+ .byte 0x44,0x0f,0x29,0x54,0x24,0x40 # movaps %xmm10,0x40(%rsp)
+ .byte 0x44,0x0f,0x29,0x5c,0x24,0x50 # movaps %xmm11,0x50(%rsp)
+ .byte 0x44,0x0f,0x29,0x64,0x24,0x60 # movaps %xmm12,0x60(%rsp)
+ .byte 0x44,0x0f,0x29,0x6c,0x24,0x70 # movaps %xmm13,0x70(%rsp)
+ .byte 0x44,0x0f,0x29,0xb4,0x24,0x80,0,0,0 # movaps %xmm14,0x80(%rsp)
+ .byte 0x44,0x0f,0x29,0xbc,0x24,0x90,0,0,0 # movaps %xmm15,0x90(%rsp)
+___
+$code.=<<___;
+ movd $power,%xmm8
+ movdqa .Linc+16(%rip),%xmm1 # 00000002000000020000000200000002
+ movdqa .Linc(%rip),%xmm0 # 00000001000000010000000000000000
+
+ pshufd \$0,%xmm8,%xmm8 # broadcast $power
+ movdqa %xmm1,%xmm7
+ movdqa %xmm1,%xmm2
+___
+########################################################################
+# calculate mask by comparing 0..15 to $power
+#
+for($i=0;$i<4;$i++) {
+$code.=<<___;
+ paddd %xmm`$i`,%xmm`$i+1`
+ pcmpeqd %xmm8,%xmm`$i`
+ movdqa %xmm7,%xmm`$i+3`
+___
+}
+for(;$i<7;$i++) {
+$code.=<<___;
+ paddd %xmm`$i`,%xmm`$i+1`
+ pcmpeqd %xmm8,%xmm`$i`
+___
+}
+$code.=<<___;
+ pcmpeqd %xmm8,%xmm7
movl \$8, %r9d
jmp .Loop_gather
.align 16
.Loop_gather:
- movl ($inp), %eax
- movl 64($inp), %r8d
+ movdqa 16*0($inp),%xmm8
+ movdqa 16*1($inp),%xmm9
+ movdqa 16*2($inp),%xmm10
+ movdqa 16*3($inp),%xmm11
+ pand %xmm0,%xmm8
+ movdqa 16*4($inp),%xmm12
+ pand %xmm1,%xmm9
+ movdqa 16*5($inp),%xmm13
+ pand %xmm2,%xmm10
+ movdqa 16*6($inp),%xmm14
+ pand %xmm3,%xmm11
+ movdqa 16*7($inp),%xmm15
leaq 128($inp), $inp
- shlq \$32, %r8
- or %r8, %rax
- movq %rax, ($out)
+ pand %xmm4,%xmm12
+ pand %xmm5,%xmm13
+ pand %xmm6,%xmm14
+ pand %xmm7,%xmm15
+ por %xmm10,%xmm8
+ por %xmm11,%xmm9
+ por %xmm12,%xmm8
+ por %xmm13,%xmm9
+ por %xmm14,%xmm8
+ por %xmm15,%xmm9
+
+ por %xmm9,%xmm8
+ pshufd \$0x4e,%xmm8,%xmm9
+ por %xmm9,%xmm8
+ movq %xmm8,($out)
leaq 8($out), $out
decl %r9d
jnz .Loop_gather
+___
+$code.=<<___ if ($win64);
+ movaps 0x00(%rsp),%xmm6
+ movaps 0x10(%rsp),%xmm7
+ movaps 0x20(%rsp),%xmm8
+ movaps 0x30(%rsp),%xmm9
+ movaps 0x40(%rsp),%xmm10
+ movaps 0x50(%rsp),%xmm11
+ movaps 0x60(%rsp),%xmm12
+ movaps 0x70(%rsp),%xmm13
+ movaps 0x80(%rsp),%xmm14
+ movaps 0x90(%rsp),%xmm15
+ add \$0xa8,%rsp
+___
+$code.=<<___;
ret
+.LSEH_end_rsaz_512_gather4:
.size rsaz_512_gather4,.-rsaz_512_gather4
+
+.align 64
+.Linc:
+ .long 0,0, 1,1
+ .long 2,2, 2,2
___
}
lea 128+24+48(%rax),%rax
+ lea .Lmul_gather4_epilogue(%rip),%rbx
+ cmp %r10,%rbx
+ jne .Lse_not_in_mul_gather4
+
+ lea 0xb0(%rax),%rax
+
+ lea -48-0xa8(%rax),%rsi
+ lea 512($context),%rdi
+ mov \$20,%ecx
+ .long 0xa548f3fc # cld; rep movsq
+
+.Lse_not_in_mul_gather4:
mov -8(%rax),%rbx
mov -16(%rax),%rbp
mov -24(%rax),%r12
pop %rdi
pop %rsi
ret
-.size sqr_handler,.-sqr_handler
+.size se_handler,.-se_handler
.section .pdata
.align 4
.rva .LSEH_end_rsaz_512_mul_by_one
.rva .LSEH_info_rsaz_512_mul_by_one
+ .rva .LSEH_begin_rsaz_512_gather4
+ .rva .LSEH_end_rsaz_512_gather4
+ .rva .LSEH_info_rsaz_512_gather4
+
.section .xdata
.align 8
.LSEH_info_rsaz_512_sqr:
.byte 9,0,0,0
.rva se_handler
.rva .Lmul_by_one_body,.Lmul_by_one_epilogue # HandlerData[]
+.LSEH_info_rsaz_512_gather4:
+ .byte 0x01,0x46,0x16,0x00
+ .byte 0x46,0xf8,0x09,0x00 #vmovaps 0x90(rsp),xmm15
+ .byte 0x3d,0xe8,0x08,0x00 #vmovaps 0x80(rsp),xmm14
+ .byte 0x34,0xd8,0x07,0x00 #vmovaps 0x70(rsp),xmm13
+ .byte 0x2e,0xc8,0x06,0x00 #vmovaps 0x60(rsp),xmm12
+ .byte 0x28,0xb8,0x05,0x00 #vmovaps 0x50(rsp),xmm11
+ .byte 0x22,0xa8,0x04,0x00 #vmovaps 0x40(rsp),xmm10
+ .byte 0x1c,0x98,0x03,0x00 #vmovaps 0x30(rsp),xmm9
+ .byte 0x16,0x88,0x02,0x00 #vmovaps 0x20(rsp),xmm8
+ .byte 0x10,0x78,0x01,0x00 #vmovaps 0x10(rsp),xmm7
+ .byte 0x0b,0x68,0x00,0x00 #vmovaps 0x00(rsp),xmm6
+ .byte 0x07,0x01,0x15,0x00 #sub rsp,0xa8
___
}