mirrors/doublecmd
2
0
Fork 0
mirror of https://github.com/doublecmd/doublecmd.git synced 2026-06-21 09:58:13 +00:00
Code Issues Projects Releases Packages Wiki Activity

FIX: SHA-256 AVX2 - use proper align

Browse source
This commit is contained in:
Alexander Koblov 2021-04-20 22:51:58 +03:00
commit 148fdbd592
1 changed files with 6 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

9
components/dcpcrypt/Hashes/sha256_avx.inc
View file

@ -57,6 +57,7 @@ procedure sha256_compress_avx(CurrentHash: PLongWord; HashBuffer: PByte; BufferC
// UNIX RDI, RSI, RDX
// WIN64: RCX, RDX, R8
asm
.balign 32
{$IF DEFINED(WIN64)}
movq %rsi, 16(%rsp)
movq %rdi, 24(%rsp)
@ -96,6 +97,7 @@ asm
shl $0x6,%rdx
lea -0x40(%rdx,%rdi,1),%rdx
mov %rdx,0x200(%rsp)
cmp %rdi,%rdx
mov (%rsi),%eax
mov 0x4(%rsi),%ebx
mov 0x8(%rsi),%ecx
@ -105,6 +107,7 @@ asm
mov 0x18(%rsi),%r10d
mov 0x1c(%rsi),%r11d
mov %rsi,0x210(%rsp)
je .LB1535
.LB7b:
lea .LK256(%rip),%rbp
@ -133,9 +136,8 @@ asm
vmovdqa %ymm9,0x40(%rsp)
vpaddd 0x60(%rbp),%ymm7,%ymm9
vmovdqa %ymm9,0x60(%rsp)
// nopl 0x0(%rax,%rax,1)
.byte 0x0f,0x1f,0x44,0x00,0x00
.balign 16
.LB100:
vpalignr $0x4,%ymm6,%ymm7,%ymm0
vpaddd %ymm4,%ymm0,%ymm0
@ -1117,8 +1119,8 @@ asm
cmp 0x200(%rsp),%rdi
ja .LB15a5
xor %rsi,%rsi
xchg %ax,%ax
.balign 16
.LB1170:
mov %edx,%edi
add 0x10(%rsp,%rsi,1),%r11d
@ -1459,6 +1461,7 @@ asm
.long 0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
.long 0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
.balign 32
.LPSHUFFLE_BYTE_FLIP_MASK:
.long 66051, 67438087, 134810123, 202182159
.long 66051, 67438087, 134810123, 202182159