Apple-2gs-SW
/
65816-crypto
mirror of https://github.com/sheumann/65816-crypto.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
65816-crypto/aes.macros

133 lines
1.7 KiB
Plaintext
Raw Normal View History

Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
macro
AddInitialRoundKey
lcla &i
.top
lda state1+&i
eor rk+&i
sta state1+&i
&i seta &i+2
aif &i<16,.top
mend
macro
NormalRound &round
aif &round/2*2=&round,.evenround
MixColumn 0,0,5,10,15,state1,state2
MixColumn 4,4,9,14,3,state1,state2
MixColumn 8,8,13,2,7,state1,state2
MixColumn 12,12,1,6,11,state1,state2
ago .done
.evenround
MixColumn 0,0,5,10,15,state2,state1
MixColumn 4,4,9,14,3,state2,state1
MixColumn 8,8,13,2,7,state2,state1
MixColumn 12,12,1,6,11,state2,state1
.done
mend
macro
MixColumn &i,&A,&B,&C,&D,&state,&out
Save a couple values on the stack rather than looking them up again. This is cycle-neutral (assuming a page-aligned DP), but it reduces instruction bytes and therefore may give a speedup on accelerators with caches.
2017-06-27 02:11:19 +00:00
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
aif (&i=0).AND.(&round<>1),.skip
Save a couple values on the stack rather than looking them up again. This is cycle-neutral (assuming a page-aligned DP), but it reduces instruction bytes and therefore may give a speedup on accelerators with caches.
2017-06-27 02:11:19 +00:00
ldy &state+&D
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
.skip
Save a couple values on the stack rather than looking them up again. This is cycle-neutral (assuming a page-aligned DP), but it reduces instruction bytes and therefore may give a speedup on accelerators with caches.
2017-06-27 02:11:19 +00:00
lda Sbox,Y
pha
Cut out a few instructions by reusing subexpressions in registers.
2017-06-27 01:53:17 +00:00
ldx &state+&A
Save a couple values on the stack rather than looking them up again. This is cycle-neutral (assuming a page-aligned DP), but it reduces instruction bytes and therefore may give a speedup on accelerators with caches.
2017-06-27 02:11:19 +00:00
eor Xtime2Sbox,X
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
ldy &state+&B
eor Xtime3Sbox,Y
Cut out a few instructions by reusing subexpressions in registers.
2017-06-27 01:53:17 +00:00
ldy &state+&C
eor Sbox,Y
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
eor rk+&round*16+&i
sta &out+&i
Cut out a few instructions by reusing subexpressions in registers.
2017-06-27 01:53:17 +00:00
Save a couple values on the stack rather than looking them up again. This is cycle-neutral (assuming a page-aligned DP), but it reduces instruction bytes and therefore may give a speedup on accelerators with caches.
2017-06-27 02:11:19 +00:00
pla
eor Xtime3Sbox,Y
Cut out a few instructions by reusing subexpressions in registers.
2017-06-27 01:53:17 +00:00
eor Sbox,X
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
ldy &state+&B
eor Xtime2Sbox,Y
eor rk+&round*16+&i+1
sta &out+&i+1
Cut out a few instructions by reusing subexpressions in registers.
2017-06-27 01:53:17 +00:00
Save a couple values on the stack rather than looking them up again. This is cycle-neutral (assuming a page-aligned DP), but it reduces instruction bytes and therefore may give a speedup on accelerators with caches.
2017-06-27 02:11:19 +00:00
lda Sbox,Y
pha
ldy &state+&D
eor Xtime3Sbox,Y
Cut out a few instructions by reusing subexpressions in registers.
2017-06-27 01:53:17 +00:00
eor Sbox,X
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
ldy &state+&C
eor Xtime2Sbox,Y
eor rk+&round*16+&i+2
sta &out+&i+2
Cut out a few instructions by reusing subexpressions in registers.
2017-06-27 01:53:17 +00:00
Save a couple values on the stack rather than looking them up again. This is cycle-neutral (assuming a page-aligned DP), but it reduces instruction bytes and therefore may give a speedup on accelerators with caches.
2017-06-27 02:11:19 +00:00
pla
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
eor Sbox,Y
Save a couple values on the stack rather than looking them up again. This is cycle-neutral (assuming a page-aligned DP), but it reduces instruction bytes and therefore may give a speedup on accelerators with caches.
2017-06-27 02:11:19 +00:00
eor Xtime3Sbox,X
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
ldy &state+&D
eor Xtime2Sbox,Y
eor rk+&round*16+&i+3
sta &out+&i+3
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
aif &i<>12,.skip2
tay
.skip2
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
mend
macro
FinalRound &round
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
FinalRoundStep 3,15,1
FinalRoundStep 15,11
FinalRoundStep 11,7
FinalRoundStep 7,3
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
FinalRoundStep 0,0
FinalRoundStep 4,4
FinalRoundStep 8,8
FinalRoundStep 12,12
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
FinalRoundStep 13,1
FinalRoundStep 1,5
FinalRoundStep 5,9
FinalRoundStep 9,13
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
FinalRoundStep 10,2
FinalRoundStep 2,10
FinalRoundStep 14,6
FinalRoundStep 6,14
mend
macro
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
FinalRoundStep &to,&from,&skipldy
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
aif C:&skipldy,.skip
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
ldy state2+&from
Save a few cycles by not reloading the last written byte between rounds.
2017-06-27 02:49:16 +00:00
.skip
Initial working implementation of AES-128 encryption.
2017-06-27 01:41:44 +00:00
lda Sbox,Y
eor rk+&round*16+&to
sta state1+&to
mend
macro
ShortRegs
sep #$30
longa off
longi off
mend
macro
LongRegs
rep #$30
longa on
longi on
mend