native.intel.common_sysV.asm Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of bcprov-lts8on Show documentation
Show all versions of bcprov-lts8on Show documentation
The Long Term Stable (LTS) Bouncy Castle Crypto package is a Java implementation of cryptographic algorithms. This jar contains the JCA/JCE provider and low-level API for the BC LTS version 2.73.6 for Java 8 and later.
The newest version!
;
; AES Key scheduler implementation in assembly
; Based upon Intel Example from the
; "Intel® Advanced Encryption Standard (AES) New Instructions Set" whitepaper
; SysV abi
global _schedule_128
global _schedule_192
global _schedule_256
global _inv_128
global _inv_192
global _inv_256
SECTION .text
align 16
;
; extern void _schedule_128(uint8_t *key (rdi), __m128i *roundKeys (rsi) );
;
_schedule_128:
movdqu xmm1,[rdi] ; rdi = user key
movdqu [rsi],xmm1 ; rsi = key schedule
ASSISTS:
aeskeygenassist xmm2,xmm1,1; $1, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [16+rsi],xmm1; 16(rsi),xmm1,
aeskeygenassist xmm2,xmm1,2; $2, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [32+rsi],xmm1; %xmm1, 32(%rsi)
aeskeygenassist xmm2,xmm1,4; $4, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [48+rsi],xmm1; %xmm1, 48(%rsi)
aeskeygenassist xmm2,xmm1,8; $8, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [64+rsi],xmm1; %xmm1, 64(%rsi)
aeskeygenassist xmm2,xmm1,16; $16, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [80+rsi],xmm1; %xmm1, 80(%rsi)
aeskeygenassist xmm2,xmm1, 32; $32, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [96+rsi],xmm1; %xmm1, 96(%rsi)
aeskeygenassist xmm2,xmm1,64; $64, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [112+rsi],xmm1; %xmm1, 112(%rsi)
aeskeygenassist xmm2,xmm1,0x80 ; $0x80, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [128+rsi],xmm1; %xmm1, 128(%rsi)
aeskeygenassist xmm2,xmm1, 0x1b; $0x1b, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [144+rsi],xmm1; %xmm1, 144(%rsi)
aeskeygenassist xmm2,xmm1,0x36; $0x36, %xmm1, %xmm2
call PREPARE_ROUNDKEY_128
movdqu [160+rsi],xmm1; %xmm1, 160(%rsi)
pxor xmm1,xmm1
pxor xmm2,xmm2
pxor xmm3,xmm3
ret
PREPARE_ROUNDKEY_128:
pshufd xmm2,xmm2,255; $255, %xmm2, %xmm2
movdqu xmm3,xmm1; %xmm1, %xmm3
pslldq xmm3,4; $4, %xmm3
pxor xmm1,xmm3; %xmm3, %xmm1
pslldq xmm3,4; $4, %xmm3
pxor xmm1,xmm3; %xmm3, %xmm1
pslldq xmm3,4; $4, %xmm3
pxor xmm1,xmm3; %xmm3, %xmm1
pxor xmm1,xmm2; %xmm2, %xmm1
ret
;
; extern void _schedule_192(uint8_t *key (rdi), __m128i *roundKeys (rsi) );
; **** Key must be within a 32 byte block ****
;
align 16
_schedule_192:
movdqu xmm1,[rdi]; (%rdi), %xmm1 rdi = user key
movdqu xmm3,[16+rdi]; 16(%rdi), %xmm3 rsi = key schedule
movdqu [rsi],xmm1; %xmm1, (%rsi)
movdqu xmm5,xmm3; %xmm3, %xmm5
aeskeygenassist xmm2,xmm3,0x01; $0x1, %xmm3, %xmm2
call PREPARE_ROUNDKEY_192
shufpd xmm5,xmm1,0; $0, %xmm1, %xmm5
movdqu [16+rsi],xmm5; %xmm5, 16(%rsi)
movdqu xmm6,xmm1; %xmm1, %xmm6
shufpd xmm6,xmm3,1; $1, %xmm3, %xmm6
movdqu [32+rsi],xmm6; %xmm6, 32(%rsi)
aeskeygenassist xmm2,xmm3,0x02; $0x2, %xmm3, %xmm2
call PREPARE_ROUNDKEY_192
movdqu [48+rsi],xmm1; %xmm1, 48(%rsi)
movdqu xmm5,xmm3; %xmm3, %xmm5
aeskeygenassist xmm2,xmm3,0x04; $0x4, %xmm3, %xmm2
call PREPARE_ROUNDKEY_192
shufpd xmm5,xmm1,0; $0, %xmm1, %xmm5
movdqu [64+rsi],xmm5; %xmm5, 64(%rsi)
movdqu xmm6,xmm1; %xmm1, %xmm6
shufpd xmm6,xmm3,1; $1, %xmm3, %xmm6
movdqu [80+rsi],xmm6; %xmm6, 80(%rsi)
aeskeygenassist xmm2,xmm3,0x08; $0x8, %xmm3, %xmm2
call PREPARE_ROUNDKEY_192
movdqu [96+rsi],xmm1; %xmm1, 96(%rsi)
movdqu xmm5,xmm3; %xmm3, %xmm5
aeskeygenassist xmm2,xmm3,0x10; $0x10, %xmm3, %xmm2
call PREPARE_ROUNDKEY_192
shufpd xmm5,xmm1,0; $0, %xmm1, %xmm5
movdqu [112+rsi],xmm5; %xmm5, 112(%rsi)
movdqu xmm6,xmm1; %xmm1, %xmm6
shufpd xmm6,xmm3,1; $1, %xmm3, %xmm6
movdqu [128+rsi],xmm6; %xmm6, 128(%rsi)
aeskeygenassist xmm2,xmm3,0x20; $0x20, %xmm3, %xmm2
call PREPARE_ROUNDKEY_192
movdqu [144+rsi],xmm1; %xmm1, 144(%rsi)
movdqu xmm5,xmm3; %xmm3, %xmm5
aeskeygenassist xmm2,xmm3,0x40; $0x40, %xmm3, %xmm2
call PREPARE_ROUNDKEY_192
shufpd xmm5,xmm1,0; $0, %xmm1, %xmm5
movdqu [160+rsi],xmm5; %xmm5, 160(%rsi)
movdqu xmm6,xmm1; %xmm1, %xmm6
shufpd xmm6,xmm3,1; $1, %xmm3, %xmm6
movdqu [176+rsi],xmm6; %xmm6, 176(%rsi)
aeskeygenassist xmm2,xmm3,0x80; $0x80, %xmm3, %xmm2
call PREPARE_ROUNDKEY_192
movdqu [192+rsi],xmm1; %xmm1, 192(%rsi)
pxor xmm1,xmm1
pxor xmm2,xmm2
pxor xmm3,xmm3
pxor xmm4,xmm4
pxor xmm5,xmm5
pxor xmm6,xmm6
ret
PREPARE_ROUNDKEY_192:
pshufd xmm2,xmm2,0x55; $0x55, %xmm2, %xmm2
movdqu xmm4,xmm1; %xmm1, %xmm4
pslldq xmm4,4; $4, %xmm4
pxor xmm1,xmm4; %xmm4, %xmm1
pslldq xmm4,4; $4, %xmm4
pxor xmm1,xmm4; %xmm4, %xmm1
pslldq xmm4,4; $4, %xmm4
pxor xmm1, xmm4; %xmm4, %xmm1
pxor xmm1,xmm2; %xmm2, %xmm1
pshufd xmm2,xmm1,0xFF; $0xff, %xmm1, %xmm2
movdqu xmm4,xmm3; %xmm3, %xmm4
pslldq xmm4,4; $4, %xmm4
pxor xmm3,xmm4; %xmm4, %xmm3
pxor xmm3,xmm2; %xmm2, %xmm3
ret
;
; extern void _schedule_192(uint8_t *key (rdi), __m128i *roundKeys (rsi) );
;
align 16
_schedule_256:
movdqu xmm1, [rdi]; (%rdi), %xmm1
movdqu xmm3,[rdi+16]; 16(%rdi), %xmm3
movdqu [rsi],xmm1; %xmm1, (%rsi)
movdqu [rsi+16],xmm3; %xmm3, 16(%rsi)
aeskeygenassist xmm2,xmm3,1; $0x1, %xmm3, %xmm2
call MAKE_RK256_a
movdqu [32+rsi],xmm1; %xmm1, 32(%rsi)
aeskeygenassist xmm2,xmm1,0; $0x0, %xmm1, %xmm2
call MAKE_RK256_b
movdqu [48+rsi],xmm3; %xmm3, 48(%rsi)
aeskeygenassist xmm2,xmm3,0x02; $0x2, %xmm3, %xmm2
call MAKE_RK256_a
movdqu [64+rsi],xmm1; %xmm1, 64(%rsi)
aeskeygenassist xmm2,xmm1,0x00; $0x0, %xmm1, %xmm2
call MAKE_RK256_b
movdqu [80+rsi],xmm3; %xmm3, 80(%rsi)
aeskeygenassist xmm2,xmm3,0x04; $0x4, %xmm3, %xmm2
call MAKE_RK256_a
movdqu [96+rsi], xmm1; %xmm1, 96(%rsi)
aeskeygenassist xmm2,xmm1,0; $0x0, %xmm1, %xmm2
call MAKE_RK256_b
movdqu [112+rsi],xmm3; %xmm3, 112(%rsi)
aeskeygenassist xmm2,xmm3,8; $0x8, %xmm3, %xmm2
call MAKE_RK256_a
movdqu [128+rsi],xmm1; %xmm1, 128(%rsi)
aeskeygenassist xmm2,xmm1,0; $0x0, %xmm1, %xmm2
call MAKE_RK256_b
movdqu [144+rsi], xmm3; %xmm3, 144(%rsi)
aeskeygenassist xmm2,xmm3,0x10; $0x10, %xmm3, %xmm2
call MAKE_RK256_a
movdqu [160+rsi],xmm1; %xmm1, 160(%rsi)
aeskeygenassist xmm2,xmm1,0; $0x0, %xmm1, %xmm2
call MAKE_RK256_b
movdqu [176+rsi],xmm3; %xmm3, 176(%rsi)
aeskeygenassist xmm2,xmm3,0x20; $0x20, %xmm3, %xmm2
call MAKE_RK256_a
movdqu [192+rsi],xmm1; %xmm1, 192(%rsi)
aeskeygenassist xmm2,xmm1,0; $0x0, %xmm1, %xmm2
call MAKE_RK256_b
movdqu [208+rsi],xmm3; %xmm3, 208(%rsi)
aeskeygenassist xmm2,xmm3,0x40; $0x40, %xmm3, %xmm2
call MAKE_RK256_a
movdqu [224+rsi], xmm1; %xmm1, 224(%rsi)
pxor xmm1,xmm1
pxor xmm2,xmm2
pxor xmm3,xmm3
pxor xmm4,xmm4
ret
MAKE_RK256_a:
pshufd xmm2,xmm2,0xFF; $0xff, %xmm2, %xmm2
movdqu xmm4,xmm1; %xmm1, %xmm4
pslldq xmm4,4; $4, %xmm4
pxor xmm1,xmm4; %xmm4, %xmm1
pslldq xmm4,4; $4, %xmm4
pxor xmm1,xmm4; %xmm4, %xmm1
pslldq xmm4,4; $4, %xmm4
pxor xmm1,xmm4; %xmm4, %xmm1
pxor xmm1,xmm2; %xmm2, %xmm1
ret
MAKE_RK256_b:
pshufd xmm2,xmm2,0xaa; $0xaa, %xmm2, %xmm2
movdqu xmm4,xmm3; %xmm3, %xmm4
pslldq xmm4,4; $4, %xmm4
pxor xmm3,xmm4; %xmm4, %xmm3
pslldq xmm4,4; $4, %xmm4
pxor xmm3,xmm4; %xmm4, %xmm3
pslldq xmm4,4; $4, %xmm4
pxor xmm3,xmm4; %xmm4, %xmm3
pxor xmm3,xmm2; %xmm2, %xmm3
ret
;
; extern void _inv_256(__m128i *roundKeys (rdi) );
;
align 16
_inv_256:
movdqu xmm1,[14*16+rdi]
movdqu xmm2,[0*16+rdi]
movdqu [0*16+rdi],xmm1
movdqu [14*16+rdi],xmm2
movdqu xmm1,[13*16+rdi]
movdqu xmm2,[1*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [1*16+rdi],xmm1
movdqu [13*16+rdi],xmm2
movdqu xmm1,[12*16+rdi]
movdqu xmm2,[2*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [2*16+rdi],xmm1
movdqu [12*16+rdi],xmm2
movdqu xmm1,[11*16+rdi]
movdqu xmm2,[3*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [3*16+rdi],xmm1
movdqu [11*16+rdi],xmm2
movdqu xmm1,[10*16+rdi]
movdqu xmm2,[4*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [4*16+rdi],xmm1
movdqu [10*16+rdi],xmm2
movdqu xmm1,[9*16+rdi]
movdqu xmm2,[5*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [5*16+rdi],xmm1
movdqu [9*16+rdi],xmm2
movdqu xmm1,[8*16+rdi]
movdqu xmm2,[6*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [6*16+rdi],xmm1
movdqu [8*16+rdi],xmm2
movdqu xmm1,[7*16+rdi]
aesimc xmm1,xmm1
movdqu [7*16+rdi],xmm1
pxor xmm1,xmm1
pxor xmm2,xmm2
ret ; inv_256
align 16
_inv_192:
movdqu xmm1,[12*16+rdi]
movdqu xmm2,[0*16+rdi]
movdqu [0*16+rdi],xmm1
movdqu [12*16+rdi],xmm2
movdqu xmm1,[11*16+rdi]
movdqu xmm2,[1*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [1*16+rdi],xmm1
movdqu [11*16+rdi],xmm2
movdqu xmm1,[10*16+rdi]
movdqu xmm2,[2*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [2*16+rdi],xmm1
movdqu [10*16+rdi],xmm2
movdqu xmm1,[9*16+rdi]
movdqu xmm2,[3*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [3*16+rdi],xmm1
movdqu [9*16+rdi],xmm2
movdqu xmm1,[8*16+rdi]
movdqu xmm2,[4*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [4*16+rdi],xmm1
movdqu [8*16+rdi],xmm2
movdqu xmm1,[7*16+rdi]
movdqu xmm2,[5*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [5*16+rdi],xmm1
movdqu [7*16+rdi],xmm2
movdqu xmm1,[6*16+rdi]
aesimc xmm1,xmm1
movdqu [6*16+rdi],xmm1
pxor xmm1,xmm1
pxor xmm2,xmm2
ret ; _inv_192
align 16
_inv_128:
movdqu xmm1,[10*16+rdi]
movdqu xmm2,[0*16+rdi]
movdqu [0*16+rdi],xmm1
movdqu [10*16+rdi],xmm2
movdqu xmm1,[9*16+rdi]
movdqu xmm2,[1*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [1*16+rdi],xmm1
movdqu [9*16+rdi],xmm2
movdqu xmm1,[8*16+rdi]
movdqu xmm2,[2*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [2*16+rdi],xmm1
movdqu [8*16+rdi],xmm2
movdqu xmm1,[7*16+rdi]
movdqu xmm2,[3*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [3*16+rdi],xmm1
movdqu [7*16+rdi],xmm2
movdqu xmm1,[6*16+rdi]
movdqu xmm2,[4*16+rdi]
aesimc xmm1,xmm1
aesimc xmm2,xmm2
movdqu [4*16+rdi],xmm1
movdqu [6*16+rdi],xmm2
movdqu xmm1,[5*16+rdi]
aesimc xmm1,xmm1
movdqu [5*16+rdi],xmm1
pxor xmm1,xmm1
pxor xmm2,xmm2
ret