+/*
+ for(i = 0; i < 5; ++i){
+ for(j = 0; j < 5; ++j){
+ a[j][i] ^= b[(4 + i) % 5][0];
+ }
+ }
+
+*/
+/* a[0..4][0]{0..7} ^= b[4][0]{0..7} */
+ sbiw XL, 5 * 8
+ sbiw r24, 40
+ rcall theta_2a
+/* a[0..4][1]{0..7} ^= b[0][0]{0..7} */
+ subi XL, lo8(4 * 5 * 8 + 8)
+ sbci XH, hi8(4 * 5 * 8 + 8)
+ rcall theta_2a
+/* a[0..4][2]{0..7} ^= b[1][0]{0..7} */
+ adiw XL, 4 * 8
+ rcall theta_2a
+/* a[0..4][3]{0..7} ^= b[2][0]{0..7} */
+ adiw XL, 4 * 8
+ rcall theta_2a
+/* a[0..4][4]{0..7} ^= b[3][0]{0..7} */
+ adiw XL, 4 * 8
+ rcall theta_2a
+/*
+ for(i = 0; i < 5; ++i){
+ for(j = 0; j < 5; ++j){
+ a[j][i] ^= rotate64_1bit_left(b[(i + 1) % 5][0]);
+ }
+ }
+*/
+/* a[0..4][0]{0..7} ^= rol(b[1][0]{0..7}) */
+ subi r24, lo8(5 * 8 - 1)
+ sbci r25, hi8(5 * 8 - 1)
+ subi XL, lo8(2 * 5 * 8 + 8)
+ sbci XH, hi8(2 * 5 * 8 + 8)
+ rcall theta_2b
+/* a[0..4][1]{0..7} ^= rol(b[2][0]{0..7}) */
+ adiw XL, 4 * 8
+ rcall theta_2b
+/* a[0..4][21]{0..7} ^= rol(b[3][0]{0..7}) */
+ adiw XL, 4 * 8
+ rcall theta_2b
+/* a[0..4][3]{0..7} ^= rol(b[4][0]{0..7}) */
+ adiw XL, 4 * 8
+ rcall theta_2b
+/* a[0..4][4]{0..7} ^= rol(b[0][0]{0..7}) */
+ subi XL, lo8(4 * 5 * 8 + 8)
+ sbci XH, hi8(4 * 5 * 8 + 8)
+ rcall theta_2b
+/*
+ -- rho & pi --
+ for(i = 0; i < 5; ++i){
+ for(j = 0; j < 5; ++j){
+ b[(2 * i + 3 * j) % 5][j] =
+ rotate64left_code(a[j][i], pgm_read_byte(&(keccak_rotate_codes[i][j])));
+ }
+ }
+
+ -- or --
+
+ const uint8_t *rot_code = (const uint8_t*)keccak_rotate_codes;
+ const uint8_t *idx_idx = (const uint8_t*)rho_pi_idx_table;
+ uint64_t *a_tmp = (uint64_t*)a;
+ for(i = 0; i < 25; ++i){
+ *((uint64_t*)(((uint8_t*)b) + pgm_read_byte(idx_idx++))) =
+ rotate64left_code(*a_tmp++, pgm_read_byte(rot_code++));
+
+ }
+
+*/
+
+.equ B_REG_L, 6
+.equ B_REG_H, 7
+
+ ldi r18, lo8(keccak_rotate_codes)
+ ldi r19, hi8(keccak_rotate_codes)
+ movw r2, r18
+ ldi r18, lo8(rho_pi_idx_table)
+ ldi r19, hi8(rho_pi_idx_table)
+ movw r4, r18
+ ldi r16, 25
+ mov r8, r16
+
+ sbiw r24, 5 * 8 + 1
+ movw YL, r24
+ sbiw XL, 8
+ movw B_REG_L, XL
+
+10:
+ ld r18, Y+
+ ld r19, Y+
+ ld r20, Y+
+ ld r21, Y+
+ ld r22, Y+
+ ld r23, Y+
+ ld r24, Y+
+ ld r25, Y+
+ movw ZL, r2
+ lpm r16, Z+
+ movw r2, ZL
+rotate64left_code:
+ ldi r30, pm_lo8(byte_rot_jmp_table)
+ ldi r31, pm_hi8(byte_rot_jmp_table)
+ mov r0, r16
+ andi r16, 0x70
+ swap r16
+ add r30, r16
+ adc r31, r1
+ mov r16, r0
+ andi r16, 0x0f
+ icall
+ clr r1
+rotate64_nbit_autodir:
+ lsr r16
+ brcc rotate64_nbit_left
+rotate64_nbit_right:
+ ldi r30, pm_lo8(rotate64_1bit_right)
+ ldi r31, pm_hi8(rotate64_1bit_right)
+ rjmp icall_r16_times
+rotate64_nbit_left:
+ ldi r30, pm_lo8(rotate64_1bit_left)
+ ldi r31, pm_hi8(rotate64_1bit_left)
+icall_r16_times:
+1: dec r16
+ brmi 2f
+ icall
+ rjmp 1b
+2:
+ movw ZL, r4
+ lpm r16, Z+
+ movw r4, ZL
+ movw XL, B_REG_L
+ add XL, r16
+ adc XH, __zero_reg__
+ st X+, r18
+ st X+, r19
+ st X+, r20
+ st X+, r21
+ st X+, r22
+ st X+, r23
+ st X+, r24
+ st X+, r25
+
+ dec r8
+ brne 10b
+/*
+ -- chi --
+ for(i = 0; i < 5; ++i){
+ a[i][0] ^= ((~(b[i][1])) & (b[i][2]));
+ a[i][1] ^= ((~(b[i][2])) & (b[i][3]));
+ a[i][2] ^= ((~(b[i][3])) & (b[i][4]));
+ a[i][3] ^= ((~(b[i][4])) & (b[i][0]));
+ a[i][4] ^= ((~(b[i][0])) & (b[i][1]));
+
+ }
+*/
+ ; memcpy(a, b, 200)
+ ; X points at b + 32 + 8 = b + 40 = b[1][0] has to point to b[0][0]
+ ldi r16, 200 / 8
+ sbiw XL, 5 * 8
+ movw ZL, XL
+ subi YL, lo8(5 * 5 * 8)
+ sbci YH, hi8(5 * 5 * 8)
+ movw r2, YL
+10:
+ .rept 8
+ ld r22, X+
+ st Y+, r22
+ .endr
+ dec r16
+ brne 10b
+
+ ; Z points at b
+ movw XL, ZL
+ movw r4, ZL
+ adiw XL, 8
+ adiw ZL, 16
+ movw YL, r2
+ ldi r18, 5
+10:
+ rcall chi_step
+ rcall chi_step
+ rcall chi_step
+ sbiw ZL, 5 * 8
+ rcall chi_step
+ sbiw XL, 5 * 8
+ rcall chi_step
+ adiw XL, 5 * 8
+ adiw ZL, 5 * 8
+ dec r18
+ brne 10b
+
+ /* -- iota -- */
+ ldi r30, lo8(keccak_rc_comp)
+ ldi r31, hi8(keccak_rc_comp)
+ add r30, r9
+ adc r31, __zero_reg__
+ lpm r20, Z+
+ movw YL, r2
+ ldi r21, 0x80
+ bst r20, 6
+ brtc 10f
+ ldd r22, Y+7
+ eor r22, r21
+ std Y+7, r22
+10:
+ bst r20, 5
+ brtc 10f
+ ldd r22, Y+3
+ eor r22, r21
+ std Y+3, r22
+10:
+ bst r20, 4
+ brtc 10f
+ ldd r22, Y+1
+ eor r22, r21
+ std Y+1, r22
+10:
+ andi r20, 0x8f
+ ld r22, Y
+ eor r22, r20
+ st Y, r22
+
+ inc r9
+ mov r16, r9
+ cpi r16, 24
+ breq 20f
+ movw r24, YL
+ movw r26, r4
+ rjmp 5b
+20:
+
+ stack_free_large3 200
+
+ pop_range 28, 29
+ pop r16
+ pop_range 2, 9
+ ret
+ .endfunc
+
+
+ .global keccak224_ctx2hash
+ .func keccak224_ctx2hash
+keccak224_ctx2hash:
+ movw r20, r22
+ ldi r22, lo8(224)
+ ldi r23, hi8(224)
+ rjmp keccak_ctx2hash
+ .endfunc
+
+ .global keccak384_ctx2hash
+ .func keccak384_ctx2hash
+keccak384_ctx2hash:
+ movw r20, r22
+ ldi r22, lo8(384)
+ ldi r23, hi8(384)
+ rjmp keccak_ctx2hash
+ .endfunc
+
+ .global keccak512_ctx2hash
+ .func keccak512_ctx2hash
+keccak512_ctx2hash:
+ movw r20, r22
+ ldi r22, lo8(512)
+ ldi r23, hi8(512)
+ rjmp keccak_ctx2hash
+ .endfunc
+
+ .global keccak256_ctx2hash
+ .func keccak256_ctx2hash
+keccak256_ctx2hash:
+ movw r20, r22
+ ldi r22, lo8(256)
+ ldi r23, hi8(256)
+ .endfunc
+
+/*
+void keccak_ctx2hash(void *dest, uint16_t length_b, keccak_ctx_t *ctx){
+ while(length_b>=ctx->r){
+ memcpy(dest, ctx->a, ctx->bs);
+ dest = (uint8_t*)dest + ctx->bs;
+ length_b -= ctx->r;
+ keccak_f1600(ctx->a);
+ }
+ memcpy(dest, ctx->a, (length_b+7)/8);
+}
+*/
+ .global keccak_ctx2hash
+ .func keccak_ctx2hash
+keccak_ctx2hash:
+ push_range 2, 10
+ movw r4, r20
+ movw r6, r24
+ movw ZL, r20
+ movw r8, r22
+ subi ZL, lo8(-ctx_r)
+ sbci ZH, hi8(-ctx_r)
+ ld r2, Z+
+ ld r3, Z+
+ ldd r10, Z+3 ; load blocksize (in bytes)
+10:
+ ; length_b = (r9:r8) ; r = (r3:r2) ; (H:L)
+ cp r2, r8
+ cpc r3, r9
+ brsh 40f
+ movw XL, r4
+ movw ZL, r6
+ mov r24, r10
+20:
+ ld r22, X+
+ st Z+, r22
+ dec r24
+ brne 20b
+ movw r6, ZL
+ sub r8, r2
+ sbc r9, r3
+ movw r24, r4
+ rcall keccak_f1600
+ rjmp 10b
+40:
+ movw XL, r4
+ movw ZL, r6
+ movw r24, r8
+ adiw r24, 7
+ lsr r25
+ ror r24
+ lsr r25
+ ror r24
+ lsr r25
+ ror r24
+ adiw r24, 0
+ breq 99f
+10:
+ ld r22, X+
+ st Z+, r22
+ sbiw r24, 1
+ brne 10b
+99:
+ pop_range 2, 10
+ ret
+ .endfunc
+
+
+ .global keccak224_init
+ .func keccak224_init
+keccak224_init:
+ movw XL, r24
+ ldi r24, lo8(1152)
+ ldi r25, hi8(1152)
+ rjmp keccak_init_1
+ .endfunc
+
+ .global keccak384_init
+ .func keccak384_init
+keccak384_init:
+ movw XL, r24
+ ldi r24, lo8( 832)
+ ldi r25, hi8( 832)
+ rjmp keccak_init_1
+ .endfunc
+
+ .global keccak512_init
+ .func keccak512_init
+keccak512_init:
+ movw XL, r24
+ ldi r24, lo8( 576)
+ ldi r25, hi8( 576)
+ rjmp keccak_init_1
+ .endfunc