Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(117)

Issues Search
    My Issues | Starred     Open | Closed | All

Unified Diff: third_party/boringssl/win-x86_64/crypto/bn/x86_64-mont5.asm

Issue 2219933002: Land BoringSSL roll on master (Closed) Base URL: git@github.com:dart-lang/sdk.git@master
Patch Set: Created 4 years, 4 months ago
Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.
Jump to:
View side-by-side diff with in-line comments
Download patch
« no previous file with comments | « third_party/boringssl/win-x86_64/crypto/bn/x86_64-mont.asm ('k') | third_party/boringssl/win-x86_64/crypto/chacha/chacha-x86_64.asm » ('j') | no next file with comments »
Expand Comments ('e') | Collapse Comments ('c') | Show Comments Hide Comments ('s')
Index: third_party/boringssl/win-x86_64/crypto/bn/x86_64-mont5.asm
diff --git a/third_party/boringssl/win-x86_64/crypto/bn/x86_64-mont5.asm b/third_party/boringssl/win-x86_64/crypto/bn/x86_64-mont5.asm
index 284318aae32ceab661f82a85e7a46d3b563b66bf..cd9a6e5d4eddf0764bf1b6385dcc102bde46662f 100644
--- a/third_party/boringssl/win-x86_64/crypto/bn/x86_64-mont5.asm
+++ b/third_party/boringssl/win-x86_64/crypto/bn/x86_64-mont5.asm
@@ -31,49 +31,151 @@ ALIGN 16
$L$mul_enter:
mov r9d,r9d
mov rax,rsp
- mov r10d,DWORD[56+rsp]
+ movd xmm5,DWORD[56+rsp]
+ lea r10,[$L$inc]
push rbx
push rbp
push r12
push r13
push r14
push r15
- lea rsp,[((-40))+rsp]
- movaps XMMWORD[rsp],xmm6
- movaps XMMWORD[16+rsp],xmm7
+
lea r11,[2+r9]
neg r11
- lea rsp,[r11*8+rsp]
+ lea rsp,[((-264))+r11*8+rsp]
and rsp,-1024
mov QWORD[8+r9*8+rsp],rax
$L$mul_body:
- mov r12,rdx
- mov r11,r10
- shr r10,3
- and r11,7
- not r10
- lea rax,[$L$magic_masks]
- and r10,3
- lea r12,[96+r11*8+r12]
- movq xmm4,QWORD[r10*8+rax]
- movq xmm5,QWORD[8+r10*8+rax]
- movq xmm6,QWORD[16+r10*8+rax]
- movq xmm7,QWORD[24+r10*8+rax]
-
- movq xmm0,QWORD[(((-96)))+r12]
- movq xmm1,QWORD[((-32))+r12]
- pand xmm0,xmm4
- movq xmm2,QWORD[32+r12]
- pand xmm1,xmm5
- movq xmm3,QWORD[96+r12]
- pand xmm2,xmm6
- por xmm0,xmm1
- pand xmm3,xmm7
+ lea r12,[128+rdx]
+ movdqa xmm0,XMMWORD[r10]
+ movdqa xmm1,XMMWORD[16+r10]
+ lea r10,[((24-112))+r9*8+rsp]
+ and r10,-16
+
+ pshufd xmm5,xmm5,0
+ movdqa xmm4,xmm1
+ movdqa xmm2,xmm1
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+DB 0x67
+ movdqa xmm3,xmm4
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[112+r10],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[128+r10],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[144+r10],xmm2
+ movdqa xmm2,xmm4
+
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[160+r10],xmm3
+ movdqa xmm3,xmm4
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[176+r10],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[192+r10],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[208+r10],xmm2
+ movdqa xmm2,xmm4
+
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[224+r10],xmm3
+ movdqa xmm3,xmm4
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[240+r10],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[256+r10],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[272+r10],xmm2
+ movdqa xmm2,xmm4
+
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[288+r10],xmm3
+ movdqa xmm3,xmm4
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[304+r10],xmm0
+
+ paddd xmm3,xmm2
+DB 0x67
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[320+r10],xmm1
+
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[336+r10],xmm2
+ pand xmm0,XMMWORD[64+r12]
+
+ pand xmm1,XMMWORD[80+r12]
+ pand xmm2,XMMWORD[96+r12]
+ movdqa XMMWORD[352+r10],xmm3
+ pand xmm3,XMMWORD[112+r12]
+ por xmm0,xmm2
+ por xmm1,xmm3
+ movdqa xmm4,XMMWORD[((-128))+r12]
+ movdqa xmm5,XMMWORD[((-112))+r12]
+ movdqa xmm2,XMMWORD[((-96))+r12]
+ pand xmm4,XMMWORD[112+r10]
+ movdqa xmm3,XMMWORD[((-80))+r12]
+ pand xmm5,XMMWORD[128+r10]
+ por xmm0,xmm4
+ pand xmm2,XMMWORD[144+r10]
+ por xmm1,xmm5
+ pand xmm3,XMMWORD[160+r10]
por xmm0,xmm2
+ por xmm1,xmm3
+ movdqa xmm4,XMMWORD[((-64))+r12]
+ movdqa xmm5,XMMWORD[((-48))+r12]
+ movdqa xmm2,XMMWORD[((-32))+r12]
+ pand xmm4,XMMWORD[176+r10]
+ movdqa xmm3,XMMWORD[((-16))+r12]
+ pand xmm5,XMMWORD[192+r10]
+ por xmm0,xmm4
+ pand xmm2,XMMWORD[208+r10]
+ por xmm1,xmm5
+ pand xmm3,XMMWORD[224+r10]
+ por xmm0,xmm2
+ por xmm1,xmm3
+ movdqa xmm4,XMMWORD[r12]
+ movdqa xmm5,XMMWORD[16+r12]
+ movdqa xmm2,XMMWORD[32+r12]
+ pand xmm4,XMMWORD[240+r10]
+ movdqa xmm3,XMMWORD[48+r12]
+ pand xmm5,XMMWORD[256+r10]
+ por xmm0,xmm4
+ pand xmm2,XMMWORD[272+r10]
+ por xmm1,xmm5
+ pand xmm3,XMMWORD[288+r10]
+ por xmm0,xmm2
+ por xmm1,xmm3
+ por xmm0,xmm1
+ pshufd xmm1,xmm0,0x4e
+ por xmm0,xmm1
lea r12,[256+r12]
- por xmm0,xmm3
-
DB 102,72,15,126,195
mov r8,QWORD[r8]
@@ -82,29 +184,14 @@ DB 102,72,15,126,195
xor r14,r14
xor r15,r15
- movq xmm0,QWORD[(((-96)))+r12]
- movq xmm1,QWORD[((-32))+r12]
- pand xmm0,xmm4
- movq xmm2,QWORD[32+r12]
- pand xmm1,xmm5
-
mov rbp,r8
mul rbx
mov r10,rax
mov rax,QWORD[rcx]
- movq xmm3,QWORD[96+r12]
- pand xmm2,xmm6
- por xmm0,xmm1
- pand xmm3,xmm7
-
imul rbp,r10
mov r11,rdx
- por xmm0,xmm2
- lea r12,[256+r12]
- por xmm0,xmm3
-
mul rbp
add r10,rax
mov rax,QWORD[8+rsi]
@@ -137,14 +224,12 @@ $L$1st_enter:
cmp r15,r9
jne NEAR $L$1st
-DB 102,72,15,126,195
add r13,rax
- mov rax,QWORD[rsi]
adc rdx,0
add r13,r11
adc rdx,0
- mov QWORD[((-16))+r15*8+rsp],r13
+ mov QWORD[((-16))+r9*8+rsp],r13
mov r13,rdx
mov r11,r10
@@ -158,33 +243,78 @@ DB 102,72,15,126,195
jmp NEAR $L$outer
ALIGN 16
$L$outer:
+ lea rdx,[((24+128))+r9*8+rsp]
+ and rdx,-16
+ pxor xmm4,xmm4
+ pxor xmm5,xmm5
+ movdqa xmm0,XMMWORD[((-128))+r12]
+ movdqa xmm1,XMMWORD[((-112))+r12]
+ movdqa xmm2,XMMWORD[((-96))+r12]
+ movdqa xmm3,XMMWORD[((-80))+r12]
+ pand xmm0,XMMWORD[((-128))+rdx]
+ pand xmm1,XMMWORD[((-112))+rdx]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[((-96))+rdx]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[((-80))+rdx]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[((-64))+r12]
+ movdqa xmm1,XMMWORD[((-48))+r12]
+ movdqa xmm2,XMMWORD[((-32))+r12]
+ movdqa xmm3,XMMWORD[((-16))+r12]
+ pand xmm0,XMMWORD[((-64))+rdx]
+ pand xmm1,XMMWORD[((-48))+rdx]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[((-32))+rdx]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[((-16))+rdx]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[r12]
+ movdqa xmm1,XMMWORD[16+r12]
+ movdqa xmm2,XMMWORD[32+r12]
+ movdqa xmm3,XMMWORD[48+r12]
+ pand xmm0,XMMWORD[rdx]
+ pand xmm1,XMMWORD[16+rdx]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[32+rdx]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[48+rdx]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[64+r12]
+ movdqa xmm1,XMMWORD[80+r12]
+ movdqa xmm2,XMMWORD[96+r12]
+ movdqa xmm3,XMMWORD[112+r12]
+ pand xmm0,XMMWORD[64+rdx]
+ pand xmm1,XMMWORD[80+rdx]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[96+rdx]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[112+rdx]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ por xmm4,xmm5
+ pshufd xmm0,xmm4,0x4e
+ por xmm0,xmm4
+ lea r12,[256+r12]
+
+ mov rax,QWORD[rsi]
+DB 102,72,15,126,195
+
xor r15,r15
mov rbp,r8
mov r10,QWORD[rsp]
- movq xmm0,QWORD[(((-96)))+r12]
- movq xmm1,QWORD[((-32))+r12]
- pand xmm0,xmm4
- movq xmm2,QWORD[32+r12]
- pand xmm1,xmm5
-
mul rbx
add r10,rax
mov rax,QWORD[rcx]
adc rdx,0
- movq xmm3,QWORD[96+r12]
- pand xmm2,xmm6
- por xmm0,xmm1
- pand xmm3,xmm7
-
imul rbp,r10
mov r11,rdx
- por xmm0,xmm2
- lea r12,[256+r12]
- por xmm0,xmm3
-
mul rbp
add r10,rax
mov rax,QWORD[8+rsi]
@@ -220,15 +350,12 @@ $L$inner_enter:
cmp r15,r9
jne NEAR $L$inner
-DB 102,72,15,126,195
-
add r13,rax
- mov rax,QWORD[rsi]
adc rdx,0
add r13,r10
- mov r10,QWORD[r15*8+rsp]
+ mov r10,QWORD[r9*8+rsp]
adc rdx,0
- mov QWORD[((-16))+r15*8+rsp],r13
+ mov QWORD[((-16))+r9*8+rsp],r13
mov r13,rdx
xor rdx,rdx
@@ -274,8 +401,7 @@ $L$copy:
mov rsi,QWORD[8+r9*8+rsp]
mov rax,1
- movaps xmm6,XMMWORD[((-88))+rsi]
- movaps xmm7,XMMWORD[((-72))+rsi]
+
mov r15,QWORD[((-48))+rsi]
mov r14,QWORD[((-40))+rsi]
mov r13,QWORD[((-32))+rsi]
@@ -312,13 +438,10 @@ DB 0x67
push r13
push r14
push r15
- lea rsp,[((-40))+rsp]
- movaps XMMWORD[rsp],xmm6
- movaps XMMWORD[16+rsp],xmm7
+
DB 0x67
- mov r10d,r9d
shl r9d,3
- shl r10d,3+2
+ lea r10,[r9*2+r9]
neg r9
@@ -328,19 +451,21 @@ DB 0x67
- lea r11,[((-64))+r9*2+rsp]
- sub r11,rsi
+
+
+ lea r11,[((-320))+r9*2+rsp]
+ sub r11,rdi
and r11,4095
cmp r10,r11
jb NEAR $L$mul4xsp_alt
sub rsp,r11
- lea rsp,[((-64))+r9*2+rsp]
+ lea rsp,[((-320))+r9*2+rsp]
jmp NEAR $L$mul4xsp_done
ALIGN 32
$L$mul4xsp_alt:
- lea r10,[((4096-64))+r9*2]
- lea rsp,[((-64))+r9*2+rsp]
+ lea r10,[((4096-320))+r9*2]
+ lea rsp,[((-320))+r9*2+rsp]
sub r11,r10
mov r10,0
cmovc r11,r10
@@ -356,8 +481,7 @@ $L$mul4x_body:
mov rsi,QWORD[40+rsp]
mov rax,1
- movaps xmm6,XMMWORD[((-88))+rsi]
- movaps xmm7,XMMWORD[((-72))+rsi]
+
mov r15,QWORD[((-48))+rsi]
mov r14,QWORD[((-40))+rsi]
mov r13,QWORD[((-32))+rsi]
@@ -375,47 +499,141 @@ $L$SEH_end_bn_mul4x_mont_gather5:
ALIGN 32
mul4x_internal:
shl r9,5
- mov r10d,DWORD[56+rax]
- lea r13,[256+r9*1+rdx]
+ movd xmm5,DWORD[56+rax]
+ lea rax,[$L$inc]
+ lea r13,[128+r9*1+rdx]
shr r9,5
- mov r11,r10
- shr r10,3
- and r11,7
- not r10
- lea rax,[$L$magic_masks]
- and r10,3
- lea r12,[96+r11*8+rdx]
- movq xmm4,QWORD[r10*8+rax]
- movq xmm5,QWORD[8+r10*8+rax]
- add r11,7
- movq xmm6,QWORD[16+r10*8+rax]
- movq xmm7,QWORD[24+r10*8+rax]
- and r11,7
-
- movq xmm0,QWORD[(((-96)))+r12]
- lea r14,[256+r12]
- movq xmm1,QWORD[((-32))+r12]
- pand xmm0,xmm4
- movq xmm2,QWORD[32+r12]
- pand xmm1,xmm5
- movq xmm3,QWORD[96+r12]
- pand xmm2,xmm6
-DB 0x67
- por xmm0,xmm1
- movq xmm1,QWORD[((-96))+r14]
-DB 0x67
- pand xmm3,xmm7
-DB 0x67
- por xmm0,xmm2
- movq xmm2,QWORD[((-32))+r14]
+ movdqa xmm0,XMMWORD[rax]
+ movdqa xmm1,XMMWORD[16+rax]
+ lea r10,[((88-112))+r9*1+rsp]
+ lea r12,[128+rdx]
+
+ pshufd xmm5,xmm5,0
+ movdqa xmm4,xmm1
+DB 0x67,0x67
+ movdqa xmm2,xmm1
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
DB 0x67
- pand xmm1,xmm4
+ movdqa xmm3,xmm4
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[112+r10],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[128+r10],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[144+r10],xmm2
+ movdqa xmm2,xmm4
+
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[160+r10],xmm3
+ movdqa xmm3,xmm4
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[176+r10],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[192+r10],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[208+r10],xmm2
+ movdqa xmm2,xmm4
+
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[224+r10],xmm3
+ movdqa xmm3,xmm4
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[240+r10],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[256+r10],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[272+r10],xmm2
+ movdqa xmm2,xmm4
+
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[288+r10],xmm3
+ movdqa xmm3,xmm4
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[304+r10],xmm0
+
+ paddd xmm3,xmm2
DB 0x67
- por xmm0,xmm3
- movq xmm3,QWORD[32+r14]
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[320+r10],xmm1
+
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[336+r10],xmm2
+ pand xmm0,XMMWORD[64+r12]
+ pand xmm1,XMMWORD[80+r12]
+ pand xmm2,XMMWORD[96+r12]
+ movdqa XMMWORD[352+r10],xmm3
+ pand xmm3,XMMWORD[112+r12]
+ por xmm0,xmm2
+ por xmm1,xmm3
+ movdqa xmm4,XMMWORD[((-128))+r12]
+ movdqa xmm5,XMMWORD[((-112))+r12]
+ movdqa xmm2,XMMWORD[((-96))+r12]
+ pand xmm4,XMMWORD[112+r10]
+ movdqa xmm3,XMMWORD[((-80))+r12]
+ pand xmm5,XMMWORD[128+r10]
+ por xmm0,xmm4
+ pand xmm2,XMMWORD[144+r10]
+ por xmm1,xmm5
+ pand xmm3,XMMWORD[160+r10]
+ por xmm0,xmm2
+ por xmm1,xmm3
+ movdqa xmm4,XMMWORD[((-64))+r12]
+ movdqa xmm5,XMMWORD[((-48))+r12]
+ movdqa xmm2,XMMWORD[((-32))+r12]
+ pand xmm4,XMMWORD[176+r10]
+ movdqa xmm3,XMMWORD[((-16))+r12]
+ pand xmm5,XMMWORD[192+r10]
+ por xmm0,xmm4
+ pand xmm2,XMMWORD[208+r10]
+ por xmm1,xmm5
+ pand xmm3,XMMWORD[224+r10]
+ por xmm0,xmm2
+ por xmm1,xmm3
+ movdqa xmm4,XMMWORD[r12]
+ movdqa xmm5,XMMWORD[16+r12]
+ movdqa xmm2,XMMWORD[32+r12]
+ pand xmm4,XMMWORD[240+r10]
+ movdqa xmm3,XMMWORD[48+r12]
+ pand xmm5,XMMWORD[256+r10]
+ por xmm0,xmm4
+ pand xmm2,XMMWORD[272+r10]
+ por xmm1,xmm5
+ pand xmm3,XMMWORD[288+r10]
+ por xmm0,xmm2
+ por xmm1,xmm3
+ por xmm0,xmm1
+ pshufd xmm1,xmm0,0x4e
+ por xmm0,xmm1
+ lea r12,[256+r12]
DB 102,72,15,126,195
- movq xmm0,QWORD[96+r14]
+
mov QWORD[((16+8))+rsp],r13
mov QWORD[((56+8))+rsp],rdi
@@ -429,26 +647,10 @@ DB 102,72,15,126,195
mov r10,rax
mov rax,QWORD[rcx]
- pand xmm2,xmm5
- pand xmm3,xmm6
- por xmm1,xmm2
-
imul rbp,r10
-
-
-
-
-
-
-
- lea r14,[((64+8))+r11*8+rsp]
+ lea r14,[((64+8))+rsp]
mov r11,rdx
- pand xmm0,xmm7
- por xmm1,xmm3
- lea r12,[512+r12]
- por xmm0,xmm1
-
mul rbp
add r10,rax
mov rax,QWORD[8+r9*1+rsi]
@@ -457,7 +659,7 @@ DB 102,72,15,126,195
mul rbx
add r11,rax
- mov rax,QWORD[16+rcx]
+ mov rax,QWORD[8+rcx]
adc rdx,0
mov r10,rdx
@@ -467,7 +669,7 @@ DB 102,72,15,126,195
adc rdx,0
add rdi,r11
lea r15,[32+r9]
- lea rcx,[64+rcx]
+ lea rcx,[32+rcx]
adc rdx,0
mov QWORD[r14],rdi
mov r13,rdx
@@ -477,7 +679,7 @@ ALIGN 32
$L$1st4x:
mul rbx
add r10,rax
- mov rax,QWORD[((-32))+rcx]
+ mov rax,QWORD[((-16))+rcx]
lea r14,[32+r14]
adc rdx,0
mov r11,rdx
@@ -493,7 +695,7 @@ $L$1st4x:
mul rbx
add r11,rax
- mov rax,QWORD[((-16))+rcx]
+ mov rax,QWORD[((-8))+rcx]
adc rdx,0
mov r10,rdx
@@ -523,7 +725,7 @@ $L$1st4x:
mul rbx
add r11,rax
- mov rax,QWORD[16+rcx]
+ mov rax,QWORD[8+rcx]
adc rdx,0
mov r10,rdx
@@ -532,7 +734,7 @@ $L$1st4x:
mov rax,QWORD[16+r15*1+rsi]
adc rdx,0
add rdi,r11
- lea rcx,[64+rcx]
+ lea rcx,[32+rcx]
adc rdx,0
mov QWORD[r14],rdi
mov r13,rdx
@@ -542,7 +744,7 @@ $L$1st4x:
mul rbx
add r10,rax
- mov rax,QWORD[((-32))+rcx]
+ mov rax,QWORD[((-16))+rcx]
lea r14,[32+r14]
adc rdx,0
mov r11,rdx
@@ -558,7 +760,7 @@ $L$1st4x:
mul rbx
add r11,rax
- mov rax,QWORD[((-16))+rcx]
+ mov rax,QWORD[((-8))+rcx]
adc rdx,0
mov r10,rdx
@@ -571,8 +773,7 @@ $L$1st4x:
mov QWORD[((-16))+r14],rdi
mov r13,rdx
-DB 102,72,15,126,195
- lea rcx,[r9*2+rcx]
+ lea rcx,[r9*1+rcx]
xor rdi,rdi
add r13,r10
@@ -583,6 +784,63 @@ DB 102,72,15,126,195
ALIGN 32
$L$outer4x:
+ lea rdx,[((16+128))+r14]
+ pxor xmm4,xmm4
+ pxor xmm5,xmm5
+ movdqa xmm0,XMMWORD[((-128))+r12]
+ movdqa xmm1,XMMWORD[((-112))+r12]
+ movdqa xmm2,XMMWORD[((-96))+r12]
+ movdqa xmm3,XMMWORD[((-80))+r12]
+ pand xmm0,XMMWORD[((-128))+rdx]
+ pand xmm1,XMMWORD[((-112))+rdx]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[((-96))+rdx]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[((-80))+rdx]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[((-64))+r12]
+ movdqa xmm1,XMMWORD[((-48))+r12]
+ movdqa xmm2,XMMWORD[((-32))+r12]
+ movdqa xmm3,XMMWORD[((-16))+r12]
+ pand xmm0,XMMWORD[((-64))+rdx]
+ pand xmm1,XMMWORD[((-48))+rdx]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[((-32))+rdx]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[((-16))+rdx]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[r12]
+ movdqa xmm1,XMMWORD[16+r12]
+ movdqa xmm2,XMMWORD[32+r12]
+ movdqa xmm3,XMMWORD[48+r12]
+ pand xmm0,XMMWORD[rdx]
+ pand xmm1,XMMWORD[16+rdx]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[32+rdx]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[48+rdx]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[64+r12]
+ movdqa xmm1,XMMWORD[80+r12]
+ movdqa xmm2,XMMWORD[96+r12]
+ movdqa xmm3,XMMWORD[112+r12]
+ pand xmm0,XMMWORD[64+rdx]
+ pand xmm1,XMMWORD[80+rdx]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[96+rdx]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[112+rdx]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ por xmm4,xmm5
+ pshufd xmm0,xmm4,0x4e
+ por xmm0,xmm4
+ lea r12,[256+r12]
+DB 102,72,15,126,195
+
mov r10,QWORD[r9*1+r14]
mov rbp,r8
mul rbx
@@ -590,25 +848,11 @@ $L$outer4x:
mov rax,QWORD[rcx]
adc rdx,0
- movq xmm0,QWORD[(((-96)))+r12]
- movq xmm1,QWORD[((-32))+r12]
- pand xmm0,xmm4
- movq xmm2,QWORD[32+r12]
- pand xmm1,xmm5
- movq xmm3,QWORD[96+r12]
-
imul rbp,r10
-DB 0x67
mov r11,rdx
mov QWORD[r14],rdi
- pand xmm2,xmm6
- por xmm0,xmm1
- pand xmm3,xmm7
- por xmm0,xmm2
lea r14,[r9*1+r14]
- lea r12,[256+r12]
- por xmm0,xmm3
mul rbp
add r10,rax
@@ -618,7 +862,7 @@ DB 0x67
mul rbx
add r11,rax
- mov rax,QWORD[16+rcx]
+ mov rax,QWORD[8+rcx]
adc rdx,0
add r11,QWORD[8+r14]
adc rdx,0
@@ -630,7 +874,7 @@ DB 0x67
adc rdx,0
add rdi,r11
lea r15,[32+r9]
- lea rcx,[64+rcx]
+ lea rcx,[32+rcx]
adc rdx,0
mov r13,rdx
jmp NEAR $L$inner4x
@@ -639,7 +883,7 @@ ALIGN 32
$L$inner4x:
mul rbx
add r10,rax
- mov rax,QWORD[((-32))+rcx]
+ mov rax,QWORD[((-16))+rcx]
adc rdx,0
add r10,QWORD[16+r14]
lea r14,[32+r14]
@@ -657,7 +901,7 @@ $L$inner4x:
mul rbx
add r11,rax
- mov rax,QWORD[((-16))+rcx]
+ mov rax,QWORD[((-8))+rcx]
adc rdx,0
add r11,QWORD[((-8))+r14]
adc rdx,0
@@ -691,7 +935,7 @@ $L$inner4x:
mul rbx
add r11,rax
- mov rax,QWORD[16+rcx]
+ mov rax,QWORD[8+rcx]
adc rdx,0
add r11,QWORD[8+r14]
adc rdx,0
@@ -702,7 +946,7 @@ $L$inner4x:
mov rax,QWORD[16+r15*1+rsi]
adc rdx,0
add rdi,r11
- lea rcx,[64+rcx]
+ lea rcx,[32+rcx]
adc rdx,0
mov QWORD[((-8))+r14],r13
mov r13,rdx
@@ -712,7 +956,7 @@ $L$inner4x:
mul rbx
add r10,rax
- mov rax,QWORD[((-32))+rcx]
+ mov rax,QWORD[((-16))+rcx]
adc rdx,0
add r10,QWORD[16+r14]
lea r14,[32+r14]
@@ -731,7 +975,7 @@ $L$inner4x:
mul rbx
add r11,rax
mov rax,rbp
- mov rbp,QWORD[((-16))+rcx]
+ mov rbp,QWORD[((-8))+rcx]
adc rdx,0
add r11,QWORD[((-8))+r14]
adc rdx,0
@@ -746,9 +990,8 @@ $L$inner4x:
mov QWORD[((-24))+r14],r13
mov r13,rdx
-DB 102,72,15,126,195
mov QWORD[((-16))+r14],rdi
- lea rcx,[r9*2+rcx]
+ lea rcx,[r9*1+rcx]
xor rdi,rdi
add r13,r10
@@ -759,16 +1002,23 @@ DB 102,72,15,126,195
cmp r12,QWORD[((16+8))+rsp]
jb NEAR $L$outer4x
+ xor rax,rax
sub rbp,r13
adc r15,r15
or rdi,r15
- xor rdi,1
+ sub rax,rdi
lea rbx,[r9*1+r14]
- lea rbp,[rdi*8+rcx]
+ mov r12,QWORD[rcx]
+ lea rbp,[rcx]
mov rcx,r9
sar rcx,3+2
mov rdi,QWORD[((56+8))+rsp]
- jmp NEAR $L$sqr4x_sub
+ dec r12
+ xor r10,r10
+ mov r13,QWORD[8+rbp]
+ mov r14,QWORD[16+rbp]
+ mov r15,QWORD[24+rbp]
+ jmp NEAR $L$sqr4x_sub_entry
global bn_power5
@@ -793,12 +1043,9 @@ $L$SEH_begin_bn_power5:
push r13
push r14
push r15
- lea rsp,[((-40))+rsp]
- movaps XMMWORD[rsp],xmm6
- movaps XMMWORD[16+rsp],xmm7
- mov r10d,r9d
+
shl r9d,3
- shl r10d,3+2
+ lea r10d,[r9*2+r9]
neg r9
mov r8,QWORD[r8]
@@ -808,19 +1055,20 @@ $L$SEH_begin_bn_power5:
- lea r11,[((-64))+r9*2+rsp]
- sub r11,rsi
+
+ lea r11,[((-320))+r9*2+rsp]
+ sub r11,rdi
and r11,4095
cmp r10,r11
jb NEAR $L$pwr_sp_alt
sub rsp,r11
- lea rsp,[((-64))+r9*2+rsp]
+ lea rsp,[((-320))+r9*2+rsp]
jmp NEAR $L$pwr_sp_done
ALIGN 32
$L$pwr_sp_alt:
- lea r10,[((4096-64))+r9*2]
- lea rsp,[((-64))+r9*2+rsp]
+ lea r10,[((4096-320))+r9*2]
+ lea rsp,[((-320))+r9*2+rsp]
sub r11,r10
mov r10,0
cmovc r11,r10
@@ -848,10 +1096,15 @@ DB 102,73,15,110,218
DB 102,72,15,110,226
call __bn_sqr8x_internal
+ call __bn_post4x_internal
call __bn_sqr8x_internal
+ call __bn_post4x_internal
call __bn_sqr8x_internal
+ call __bn_post4x_internal
call __bn_sqr8x_internal
+ call __bn_post4x_internal
call __bn_sqr8x_internal
+ call __bn_post4x_internal
DB 102,72,15,126,209
DB 102,72,15,126,226
@@ -1397,9 +1650,9 @@ DB 0x67
mov QWORD[((-16))+rdi],rbx
mov QWORD[((-8))+rdi],r8
DB 102,72,15,126,213
-sqr8x_reduction:
+__bn_sqr8x_reduction:
xor rax,rax
- lea rcx,[r9*2+rbp]
+ lea rcx,[rbp*1+r9]
lea rdx,[((48+8))+r9*2+rsp]
mov QWORD[((0+8))+rsp],rcx
lea rdi,[((48+8))+r9*1+rsp]
@@ -1432,14 +1685,14 @@ DB 0x67
ALIGN 32
$L$8x_reduce:
mul rbx
- mov rax,QWORD[16+rbp]
+ mov rax,QWORD[8+rbp]
neg r8
mov r8,rdx
adc r8,0
mul rbx
add r9,rax
- mov rax,QWORD[32+rbp]
+ mov rax,QWORD[16+rbp]
adc rdx,0
add r8,r9
mov QWORD[((48-8+8))+rcx*8+rsp],rbx
@@ -1448,7 +1701,7 @@ $L$8x_reduce:
mul rbx
add r10,rax
- mov rax,QWORD[48+rbp]
+ mov rax,QWORD[24+rbp]
adc rdx,0
add r9,r10
mov rsi,QWORD[((32+8))+rsp]
@@ -1457,7 +1710,7 @@ $L$8x_reduce:
mul rbx
add r11,rax
- mov rax,QWORD[64+rbp]
+ mov rax,QWORD[32+rbp]
adc rdx,0
imul rsi,r8
add r10,r11
@@ -1466,7 +1719,7 @@ $L$8x_reduce:
mul rbx
add r12,rax
- mov rax,QWORD[80+rbp]
+ mov rax,QWORD[40+rbp]
adc rdx,0
add r11,r12
mov r12,rdx
@@ -1474,7 +1727,7 @@ $L$8x_reduce:
mul rbx
add r13,rax
- mov rax,QWORD[96+rbp]
+ mov rax,QWORD[48+rbp]
adc rdx,0
add r12,r13
mov r13,rdx
@@ -1482,7 +1735,7 @@ $L$8x_reduce:
mul rbx
add r14,rax
- mov rax,QWORD[112+rbp]
+ mov rax,QWORD[56+rbp]
adc rdx,0
add r13,r14
mov r14,rdx
@@ -1500,7 +1753,7 @@ $L$8x_reduce:
dec ecx
jnz NEAR $L$8x_reduce
- lea rbp,[128+rbp]
+ lea rbp,[64+rbp]
xor rax,rax
mov rdx,QWORD[((8+8))+rsp]
cmp rbp,QWORD[((0+8))+rsp]
@@ -1526,14 +1779,14 @@ ALIGN 32
$L$8x_tail:
mul rbx
add r8,rax
- mov rax,QWORD[16+rbp]
+ mov rax,QWORD[8+rbp]
mov QWORD[rdi],r8
mov r8,rdx
adc r8,0
mul rbx
add r9,rax
- mov rax,QWORD[32+rbp]
+ mov rax,QWORD[16+rbp]
adc rdx,0
add r8,r9
lea rdi,[8+rdi]
@@ -1542,7 +1795,7 @@ $L$8x_tail:
mul rbx
add r10,rax
- mov rax,QWORD[48+rbp]
+ mov rax,QWORD[24+rbp]
adc rdx,0
add r9,r10
mov r10,rdx
@@ -1550,7 +1803,7 @@ $L$8x_tail:
mul rbx
add r11,rax
- mov rax,QWORD[64+rbp]
+ mov rax,QWORD[32+rbp]
adc rdx,0
add r10,r11
mov r11,rdx
@@ -1558,7 +1811,7 @@ $L$8x_tail:
mul rbx
add r12,rax
- mov rax,QWORD[80+rbp]
+ mov rax,QWORD[40+rbp]
adc rdx,0
add r11,r12
mov r12,rdx
@@ -1566,7 +1819,7 @@ $L$8x_tail:
mul rbx
add r13,rax
- mov rax,QWORD[96+rbp]
+ mov rax,QWORD[48+rbp]
adc rdx,0
add r12,r13
mov r13,rdx
@@ -1574,7 +1827,7 @@ $L$8x_tail:
mul rbx
add r14,rax
- mov rax,QWORD[112+rbp]
+ mov rax,QWORD[56+rbp]
adc rdx,0
add r13,r14
mov r14,rdx
@@ -1592,7 +1845,7 @@ $L$8x_tail:
dec ecx
jnz NEAR $L$8x_tail
- lea rbp,[128+rbp]
+ lea rbp,[64+rbp]
mov rdx,QWORD[((8+8))+rsp]
cmp rbp,QWORD[((0+8))+rsp]
jae NEAR $L$8x_tail_done
@@ -1616,6 +1869,15 @@ $L$8x_tail:
ALIGN 32
$L$8x_tail_done:
add r8,QWORD[rdx]
+ adc r9,0
+ adc r10,0
+ adc r11,0
+ adc r12,0
+ adc r13,0
+ adc r14,0
+ adc r15,0
+
+
xor rax,rax
neg rsi
@@ -1629,7 +1891,7 @@ $L$8x_no_tail:
adc r14,QWORD[48+rdi]
adc r15,QWORD[56+rdi]
adc rax,0
- mov rcx,QWORD[((-16))+rbp]
+ mov rcx,QWORD[((-8))+rbp]
xor rsi,rsi
DB 102,72,15,126,213
@@ -1647,40 +1909,58 @@ DB 102,73,15,126,217
cmp rdi,rdx
jb NEAR $L$8x_reduction_loop
+ DB 0F3h,0C3h ;repret
+
- sub rcx,r15
+ALIGN 32
+__bn_post4x_internal:
+ mov r12,QWORD[rbp]
lea rbx,[r9*1+rdi]
- adc rsi,rsi
mov rcx,r9
- or rax,rsi
DB 102,72,15,126,207
- xor rax,1
+ neg rax
DB 102,72,15,126,206
- lea rbp,[rax*8+rbp]
sar rcx,3+2
- jmp NEAR $L$sqr4x_sub
+ dec r12
+ xor r10,r10
+ mov r13,QWORD[8+rbp]
+ mov r14,QWORD[16+rbp]
+ mov r15,QWORD[24+rbp]
+ jmp NEAR $L$sqr4x_sub_entry
-ALIGN 32
+ALIGN 16
$L$sqr4x_sub:
-DB 0x66
- mov r12,QWORD[rbx]
- mov r13,QWORD[8+rbx]
- sbb r12,QWORD[rbp]
- mov r14,QWORD[16+rbx]
- sbb r13,QWORD[16+rbp]
- mov r15,QWORD[24+rbx]
- lea rbx,[32+rbx]
- sbb r14,QWORD[32+rbp]
+ mov r12,QWORD[rbp]
+ mov r13,QWORD[8+rbp]
+ mov r14,QWORD[16+rbp]
+ mov r15,QWORD[24+rbp]
+$L$sqr4x_sub_entry:
+ lea rbp,[32+rbp]
+ not r12
+ not r13
+ not r14
+ not r15
+ and r12,rax
+ and r13,rax
+ and r14,rax
+ and r15,rax
+
+ neg r10
+ adc r12,QWORD[rbx]
+ adc r13,QWORD[8+rbx]
+ adc r14,QWORD[16+rbx]
+ adc r15,QWORD[24+rbx]
mov QWORD[rdi],r12
- sbb r15,QWORD[48+rbp]
- lea rbp,[64+rbp]
+ lea rbx,[32+rbx]
mov QWORD[8+rdi],r13
+ sbb r10,r10
mov QWORD[16+rdi],r14
mov QWORD[24+rdi],r15
lea rdi,[32+rdi]
inc rcx
jnz NEAR $L$sqr4x_sub
+
mov r10,r9
neg r9
DB 0F3h,0C3h ;repret
@@ -1718,13 +1998,9 @@ DB 0x67
push r13
push r14
push r15
- lea rsp,[((-40))+rsp]
- movaps XMMWORD[rsp],xmm6
- movaps XMMWORD[16+rsp],xmm7
-DB 0x67
- mov r10d,r9d
+
shl r9d,3
- shl r10d,3+2
+ lea r10,[r9*2+r9]
neg r9
mov r8,QWORD[r8]
@@ -1734,19 +2010,20 @@ DB 0x67
- lea r11,[((-64))+r9*2+rsp]
- sub r11,rsi
+
+ lea r11,[((-320))+r9*2+rsp]
+ sub r11,rdi
and r11,4095
cmp r10,r11
jb NEAR $L$from_sp_alt
sub rsp,r11
- lea rsp,[((-64))+r9*2+rsp]
+ lea rsp,[((-320))+r9*2+rsp]
jmp NEAR $L$from_sp_done
ALIGN 32
$L$from_sp_alt:
- lea r10,[((4096-64))+r9*2]
- lea rsp,[((-64))+r9*2+rsp]
+ lea r10,[((4096-320))+r9*2]
+ lea rsp,[((-320))+r9*2+rsp]
sub r11,r10
mov r10,0
cmovc r11,r10
@@ -1797,7 +2074,8 @@ DB 102,72,15,110,209
DB 0x67
mov rbp,rcx
DB 102,73,15,110,218
- call sqr8x_reduction
+ call __bn_sqr8x_reduction
+ call __bn_post4x_internal
pxor xmm0,xmm0
lea rax,[48+rsp]
@@ -1847,55 +2125,171 @@ $L$scatter_epilogue:
global bn_gather5
-ALIGN 16
+ALIGN 32
bn_gather5:
$L$SEH_begin_bn_gather5:
-DB 0x48,0x83,0xec,0x28
-DB 0x0f,0x29,0x34,0x24
-DB 0x0f,0x29,0x7c,0x24,0x10
- mov r11d,r9d
- shr r9d,3
- and r11,7
- not r9d
- lea rax,[$L$magic_masks]
- and r9d,3
- lea r8,[128+r11*8+r8]
- movq xmm4,QWORD[r9*8+rax]
- movq xmm5,QWORD[8+r9*8+rax]
- movq xmm6,QWORD[16+r9*8+rax]
- movq xmm7,QWORD[24+r9*8+rax]
+DB 0x4c,0x8d,0x14,0x24
+DB 0x48,0x81,0xec,0x08,0x01,0x00,0x00
+ lea rax,[$L$inc]
+ and rsp,-16
+
+ movd xmm5,r9d
+ movdqa xmm0,XMMWORD[rax]
+ movdqa xmm1,XMMWORD[16+rax]
+ lea r11,[128+r8]
+ lea rax,[128+rsp]
+
+ pshufd xmm5,xmm5,0
+ movdqa xmm4,xmm1
+ movdqa xmm2,xmm1
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa xmm3,xmm4
+
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[(-128)+rax],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[(-112)+rax],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[(-96)+rax],xmm2
+ movdqa xmm2,xmm4
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[(-80)+rax],xmm3
+ movdqa xmm3,xmm4
+
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[(-64)+rax],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[(-48)+rax],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[(-32)+rax],xmm2
+ movdqa xmm2,xmm4
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[(-16)+rax],xmm3
+ movdqa xmm3,xmm4
+
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[rax],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[16+rax],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[32+rax],xmm2
+ movdqa xmm2,xmm4
+ paddd xmm1,xmm0
+ pcmpeqd xmm0,xmm5
+ movdqa XMMWORD[48+rax],xmm3
+ movdqa xmm3,xmm4
+
+ paddd xmm2,xmm1
+ pcmpeqd xmm1,xmm5
+ movdqa XMMWORD[64+rax],xmm0
+ movdqa xmm0,xmm4
+
+ paddd xmm3,xmm2
+ pcmpeqd xmm2,xmm5
+ movdqa XMMWORD[80+rax],xmm1
+ movdqa xmm1,xmm4
+
+ paddd xmm0,xmm3
+ pcmpeqd xmm3,xmm5
+ movdqa XMMWORD[96+rax],xmm2
+ movdqa xmm2,xmm4
+ movdqa XMMWORD[112+rax],xmm3
jmp NEAR $L$gather
-ALIGN 16
-$L$gather:
- movq xmm0,QWORD[(((-128)))+r8]
- movq xmm1,QWORD[((-64))+r8]
- pand xmm0,xmm4
- movq xmm2,QWORD[r8]
- pand xmm1,xmm5
- movq xmm3,QWORD[64+r8]
- pand xmm2,xmm6
- por xmm0,xmm1
- pand xmm3,xmm7
-DB 0x67,0x67
- por xmm0,xmm2
- lea r8,[256+r8]
- por xmm0,xmm3
+ALIGN 32
+$L$gather:
+ pxor xmm4,xmm4
+ pxor xmm5,xmm5
+ movdqa xmm0,XMMWORD[((-128))+r11]
+ movdqa xmm1,XMMWORD[((-112))+r11]
+ movdqa xmm2,XMMWORD[((-96))+r11]
+ pand xmm0,XMMWORD[((-128))+rax]
+ movdqa xmm3,XMMWORD[((-80))+r11]
+ pand xmm1,XMMWORD[((-112))+rax]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[((-96))+rax]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[((-80))+rax]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[((-64))+r11]
+ movdqa xmm1,XMMWORD[((-48))+r11]
+ movdqa xmm2,XMMWORD[((-32))+r11]
+ pand xmm0,XMMWORD[((-64))+rax]
+ movdqa xmm3,XMMWORD[((-16))+r11]
+ pand xmm1,XMMWORD[((-48))+rax]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[((-32))+rax]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[((-16))+rax]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[r11]
+ movdqa xmm1,XMMWORD[16+r11]
+ movdqa xmm2,XMMWORD[32+r11]
+ pand xmm0,XMMWORD[rax]
+ movdqa xmm3,XMMWORD[48+r11]
+ pand xmm1,XMMWORD[16+rax]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[32+rax]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[48+rax]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ movdqa xmm0,XMMWORD[64+r11]
+ movdqa xmm1,XMMWORD[80+r11]
+ movdqa xmm2,XMMWORD[96+r11]
+ pand xmm0,XMMWORD[64+rax]
+ movdqa xmm3,XMMWORD[112+r11]
+ pand xmm1,XMMWORD[80+rax]
+ por xmm4,xmm0
+ pand xmm2,XMMWORD[96+rax]
+ por xmm5,xmm1
+ pand xmm3,XMMWORD[112+rax]
+ por xmm4,xmm2
+ por xmm5,xmm3
+ por xmm4,xmm5
+ lea r11,[256+r11]
+ pshufd xmm0,xmm4,0x4e
+ por xmm0,xmm4
movq QWORD[rcx],xmm0
lea rcx,[8+rcx]
sub edx,1
jnz NEAR $L$gather
- movaps xmm6,XMMWORD[rsp]
- movaps xmm7,XMMWORD[16+rsp]
- lea rsp,[40+rsp]
+
+ lea rsp,[r10]
DB 0F3h,0C3h ;repret
$L$SEH_end_bn_gather5:
ALIGN 64
-$L$magic_masks:
- DD 0,0,0,0,0,0,-1,-1
- DD 0,0,0,0,0,0,0,0
+$L$inc:
+ DD 0,0,1,1
+ DD 2,2,2,2
DB 77,111,110,116,103,111,109,101,114,121,32,77,117,108,116,105
DB 112,108,105,99,97,116,105,111,110,32,119,105,116,104,32,115
DB 99,97,116,116,101,114,47,103,97,116,104,101,114,32,102,111
@@ -1937,19 +2331,16 @@ mul_handler:
lea r10,[$L$mul_epilogue]
cmp rbx,r10
- jb NEAR $L$body_40
+ ja NEAR $L$body_40
mov r10,QWORD[192+r8]
mov rax,QWORD[8+r10*8+rax]
+
jmp NEAR $L$body_proceed
$L$body_40:
mov rax,QWORD[40+rax]
$L$body_proceed:
-
- movaps xmm0,XMMWORD[((-88))+rax]
- movaps xmm1,XMMWORD[((-72))+rax]
-
mov rbx,QWORD[((-8))+rax]
mov rbp,QWORD[((-16))+rax]
mov r12,QWORD[((-24))+rax]
@@ -1962,8 +2353,6 @@ $L$body_proceed:
mov QWORD[224+r8],r13
mov QWORD[232+r8],r14
mov QWORD[240+r8],r15
- movups XMMWORD[512+r8],xmm0
- movups XMMWORD[528+r8],xmm1
$L$common_seh_tail:
mov rdi,QWORD[8+rax]
@@ -2049,8 +2438,7 @@ DB 9,0,0,0
DD $L$from_body wrt ..imagebase,$L$from_epilogue wrt ..imagebase
ALIGN 8
$L$SEH_info_bn_gather5:
-DB 0x01,0x0d,0x05,0x00
-DB 0x0d,0x78,0x01,0x00
-DB 0x08,0x68,0x00,0x00
-DB 0x04,0x42,0x00,0x00
+DB 0x01,0x0b,0x03,0x0a
+DB 0x0b,0x01,0x21,0x00
+DB 0x04,0xa3,0x00,0x00
ALIGN 8
« no previous file with comments | « third_party/boringssl/win-x86_64/crypto/bn/x86_64-mont.asm ('k') | third_party/boringssl/win-x86_64/crypto/chacha/chacha-x86_64.asm » ('j') | no next file with comments »

Powered by Google App Engine
This is Rietveld 408576698