X-Git-Url: https://git.cryptolib.org/?a=blobdiff_plain;f=bmw%2Fbmw_small-tinyasm.S;h=fca301a39e093fa83d89ba06059d4e1697c55352;hb=709c6948180356089680fd243927b22133ba64d0;hp=198aa7024688413f0bcb985079237e80bb3532b1;hpb=03f98ddd20eaabe1cfb0e4933ed559295543a55b;p=avr-crypto-lib.git diff --git a/bmw/bmw_small-tinyasm.S b/bmw/bmw_small-tinyasm.S index 198aa70..fca301a 100644 --- a/bmw/bmw_small-tinyasm.S +++ b/bmw/bmw_small-tinyasm.S @@ -109,13 +109,22 @@ s2: .byte 2, 1,12,25 s3: .byte 2, 2,15,29 s4: .byte 1, 0, 0, 0 s5: .byte 2, 0, 0, 0 - -eor_r22_in_r16: - eor r16, r22 - eor r17, r23 - eor r18, r24 - eor r19, r25 - ret +/* +s0: .byte 0x34, 19 +s1: .byte 0x28, 23 +s2: .byte 0x9C, 25 +s3: .byte 0xAF, 29 +s4: .byte 0x00, 0 +s5: .byte 0x80, 0 +*/ +acc2 = 8 +acc3 = 9 +h0 = 10 +h1 = 11 +m0 = 12 +m1 = 13 +acc0 = 14 +acc1 = 15 /* param x: r22:r23:r24:25 @@ -123,8 +132,10 @@ eor_r22_in_r16: */ sn: push_range 2, 5 - push r17 - push r19 + push acc0 + push acc1 + push acc2 + push acc3 ldi r30, lo8(s_table) ldi r31, hi8(s_table) lsl r20 @@ -135,31 +146,32 @@ sn: movw r4, r24 lpm r20, Z+ rcall shiftright32 - movw r16, r22 - movw r18, r24 + movw acc0, r22 + movw acc2, r24 ;--- movw r22, r2 movw r24, r4 lpm r20, Z+ rcall shiftleft32 - rcall eor_r22_in_r16 + rcall eor32_to_acc ;--- movw r22, r2 movw r24, r4 lpm r20, Z+ rcall rotateleft32 - rcall eor_r22_in_r16 + rcall eor32_to_acc ;--- movw r22, r2 movw r24, r4 lpm r20, Z+ rcall rotateleft32 - eor r22, r16 - eor r23, r17 - eor r24, r18 - eor r25, r19 - pop r19 - pop r17 + rcall eor32_to_acc + movw r22, acc0 + movw r24, acc2 + pop acc3 + pop acc2 + pop acc1 + pop acc0 pop_range 2, 5 ret @@ -172,6 +184,7 @@ sn: memxor_short: ; tst r20 ; breq memxor_exit + ldi r20, 64 10: ld r21, X ld r22, Z+ eor r21, r22 @@ -189,42 +202,6 @@ h1 = 5 m0 = 6 m1 = 7 -add_hx_to_w: - movw r26, h0 - add r26, r16 - adc r27, r1 - ld r22, Y - ldd r23, Y+1 - ldd r24, Y+2 - ldd r25, Y+3 - lsl r20 - rol r21 - brcs 30f - /* addition */ - ld r0, X+ - add r22, r0 - ld r0, X+ - adc r23, r0 - ld r0, X+ - adc r24, r0 - ld r0, X+ - adc r25, r0 - rjmp 50f -30: /* substract */ - ld r0, X+ - sub r22, r0 - ld r0, X+ - sbc r23, r0 - ld r0, X+ - sbc r24, r0 - ld r0, X+ - sbc r25, r0 -50: - st Y+, r22 - st Y+, r23 - st Y+, r24 - st Y+, r25 - ret /******************************************************************************/ load32_from_X: @@ -241,6 +218,13 @@ load32_from_Y: ld r25, Y+ ret +store32_to_Y: + st Y+, r22 + st Y+, r23 + st Y+, r24 + st Y+, r25 + ret + add_X_to_32: ld r0, X+ add r22, r0 @@ -251,6 +235,21 @@ add_X_to_32: ld r0, X+ adc r25, r0 ret + +store_acc_to_dec_X: + st -X, acc3 + st -X, acc2 + st -X, acc1 + st -X, acc0 + ret + +store32_to_X: + st X+, r22 + st X+, r23 + st X+, r24 + st X+, r25 + ret + /******************************************************************************/ /* param q: r28:r29 (Y) @@ -259,142 +258,13 @@ add_X_to_32: */ f0_hacktable: - .byte 0x03, 0x11 - .byte 0xDD, 0xB3 - .byte 0x2A, 0x79 - .byte 0x07, 0xAA - .byte 0x51, 0xC2 -f0_indextable: - .byte 5*4,7*4,10*4,13*4,14*4 -; .byte 0 ; just for alignment -f0_s_table: - .byte 0,1,2,3,4 - .byte 0,1,2,3,4 - .byte 0,1,2,3,4 -; .byte 0 - -f0: - movw h0, r26 - movw q0, r28 - movw m0, r30 -;--- DBG -; push_range 22, 25 -; movw r24, r26 -; ldi r22, 'H' -; rcall printX -; pop_range 22, 25 -;--- END DBG -;--- DBG -; push_range 22, 25 -; movw r24, r30 -; ldi r22, 'M' -; rcall printX -; pop_range 22, 25 -;--- END DBG - /* xor m into h */ - ldi r20, 64 - rcall memxor_short - movw r30, m0 - movw r26, h0 - - /* set q to zero */ - ldi r22, 64 -10: st Y+, r1 - dec r22 - brne 10b - movw r28, q0 - /* calculate W and store it in Q */ - ldi r19, 5 -30: - ldi r18, 16 - /* load initial index */ - ldi r30, lo8(f0_indextable-1) - ldi r31, hi8(f0_indextable-1) - add r30, r19 - adc r31, r1 - lpm r16, Z - /* load values from hacktable */ - ldi r30, lo8(f0_hacktable-2) - ldi r31, hi8(f0_hacktable-2) - lsl r19 - add r30, r19 - adc r31, r1 - lsr r19 - lpm r21, Z+ - lpm r20, Z -40: - call add_hx_to_w - subi r16, -4 - andi r16, 0x0f<<2 - dec r18 - brne 40b - movw r28, q0 - dec r19 - brne 30b - movw r26, h0 -;--- DBG -; push_range 22, 25 -; movw r24, r28 -; ldi r22, 'W' -; rcall printX -; pop_range 22, 25 -;--- END DBG - /* xor m into h */ - ldi r20, 64 - movw r26, h0 - movw r30, m0 - rcall memxor_short - sbiw r26, 60 -;--- - ldi r30, lo8(f0_s_table) - ldi r31, hi8(f0_s_table) - ldi r21, 15 - mov r8, r21 -50: - ldd r22, Y+0 - ldd r23, Y+1 - ldd r24, Y+2 - ldd r25, Y+3 - lpm r20, Z+ - movw r2, r30 - rcall sn - movw r30, r2 - - rcall add_X_to_32 - - st Y+, r22 - st Y+, r23 - st Y+, r24 - st Y+, r25 - dec r8 - brne 50b -;--- - ldd r22, Y+0 - ldd r23, Y+1 - ldd r24, Y+2 - ldd r25, Y+3 - clr r20 - rcall sn - movw r30, r2 - movw r26, h0 - rcall add_X_to_32 - sbiw r26, 4 - std Y+0, r22 - std Y+1, r23 - std Y+2, r24 - std Y+3, r25 - sbiw r28, 15*4 - movw r20, h0 - movw r22, m0 - ret - -/******************************************************************************/ + .byte 0x03, 0x11, 5*4 + .byte 0xDD, 0xB3, 7*4 + .byte 0x2A, 0x79, 10*4 + .byte 0x07, 0xAA, 13*4 + .byte 0x51, 0xC2, 14*4 + .byte 0 ; just for alignment -const_lut: - .long 0x55555550, 0x5aaaaaa5, 0x5ffffffa, 0x6555554f - .long 0x6aaaaaa4, 0x6ffffff9, 0x7555554e, 0x7aaaaaa3 - .long 0x7ffffff8, 0x8555554d, 0x8aaaaaa2, 0x8ffffff7 - .long 0x9555554c, 0x9aaaaaa1, 0x9ffffff6, 0xa555554b /******************************************************************************* * uint32_t addelment(uint8_t j, const uint32_t* m, const uint32_t* h){ @@ -464,10 +334,7 @@ load_rotate_add_M: movw r26, m0 add r26, r0 adc r27, r1 - ld r22, X+ - ld r23, X+ - ld r24, X+ - ld r25, X+ + rcall load32_from_X inc r20 rcall rotateleft32 brts 10f @@ -483,17 +350,16 @@ addelement: mov j, r24 movw h0, r20 movw m0, r22 - lsl r24 - lsl r24 - mov r28, r24 - ldi r30, lo8(const_lut) - ldi r31, hi8(const_lut) - add r30, r24 - adc r31, r1 - lpm acc0, Z+ - lpm acc1, Z+ - lpm acc2, Z+ - lpm acc3, Z+ + sbiw r26, 4 + rcall load_acc_from_X + ldi r24, 0x55 + add acc0, r24 + adc acc1, r24 + adc acc2, r24 + ldi r24, 5 + adc acc3, r24 + rcall store_acc_to_dec_X + adiw r26, 4 clt mov r20, j rcall load_rotate_add_M @@ -511,18 +377,18 @@ addelement: movw r26, h0 add r26, j adc r27, r1 - ld r0, X+ - eor acc0, r0 - ld r0, X+ - eor acc1, r0 - ld r0, X+ - eor acc2, r0 - ld r0, X+ - eor acc3, r0 + rcall load32_from_X + rcall eor32_to_acc ;--- ret /******************************************************************************/ +load_sn_add: + rcall load32_from_X + rcall sn + rcall add32_to_acc + ret + /* param q: r26:r27 param m: r22:r23 @@ -545,21 +411,14 @@ expand1: rcall expand_intro ldi r19, 1 10: - rcall load32_from_X mov r20, r19 andi r20, 3 - rcall sn - rcall add32_to_acc + rcall load_sn_add inc r19 cpi r19, 17 brne 10b -expand1_exit: -; adiw r26, 63 - st X+, acc0 - st X+, acc1 - st X+, acc2 - st X+, acc3 - ret + rjmp expand2_exit + /******************************************************************************/ /* @@ -569,8 +428,14 @@ expand1_exit: param j: r24 */ +f2_1_shift_table: + .byte 0x2B, 0x64, 0x66, 0x03, 0x51, 0x55, 0x87, 0x55 +f2_2_shift_table: + .byte (2<<1), (7<<1), (4<<1), (3<<1), (4<<1)+1, (6<<1)+1, (6<<1) + expand2_rot_table: - .byte 0,3,0,7,0,13,0,16,0,19,0,23,0,27 + .byte 3,7,13,16,19,23,27 +; .byte 0 ; just for alignment expand2: rcall expand_intro @@ -579,22 +444,21 @@ expand2: ldi r31, hi8(expand2_rot_table) 10: rcall load32_from_X - mov r20, r19 + sbrs r19, 0 + rjmp 12f lpm r20, Z+ rcall rotateleft32 - rcall add32_to_acc +12: rcall add32_to_acc dec r19 brne 10b - rcall load32_from_X ldi r20, 4 - rcall sn - rcall add32_to_acc - rcall load32_from_X + rcall load_sn_add ldi r20, 5 - rcall sn - rcall add32_to_acc - - rjmp expand1_exit + rcall load_sn_add +expand2_exit: + adiw r26, 4 + rcall store_acc_to_dec_X + ret /******************************************************************************/ /* @@ -608,13 +472,185 @@ expand2: param h: r20:r21 param j: r24 */ + +/******************************************************************************/ +/* + param q: r24:r25 + param m: r22:r23 + param h: r20:r21 +*/ + +/******************************************************************************/ +/* + param ctx: r24:r25 + param msg: r22:r23 +*/ +/* f0 + param q: r28:r29 (Y) + param h: r26:r27 (X) + param m: r30:r31 (Z) +*/ +/* f1 + param q: r24:r25 + param m: r22:r23 + param h: r20:r21 +*/ +/* f2 + param q: r24:r25 + param m: r22:r23 + param h: r20:r21 +*/ +q0 = 2 +q1 = 3 +h0 = 4 +h1 = 5 +m0 = 6 +m1 = 7 + + +.global bmw_small_nextBlock +.global bmw224_nextBlock +.global bmw256_nextBlock +bmw_small_nextBlock: +bmw224_nextBlock: +bmw256_nextBlock: + push_range 28, 29 + push_range 2, 17 + stack_alloc_large 32*4, r28, r29 + ldi r16, 0x4f + push r16 + ldi r16, 0xff + push r16 + push r16 + ldi r16, 0xfb + push r16 + adiw r28, 1 +; push_range 28, 29 /* push Q */ +; push_range 22, 25 /* push M & H */ + /* increment counter */ + movw r26, r24 + movw r2, r26 + adiw r26, 63 + adiw r26, 1 + rcall load_acc_from_X + ldi r19, 1 + add acc0, r19 + adc acc1, r1 + adc acc2, r1 + adc acc3, r1 + rcall store_acc_to_dec_X + /* call f0 */ + movw r30, r22 + movw r26, r24 +f0: + movw h0, r26 + movw q0, r28 + movw m0, r30 + /* xor m into h */ +; ldi r20, 64 + rcall memxor_short + movw r30, m0 + movw r26, h0 + + /* set q to zero */ + ldi r22, 64 +10: st Y+, r1 + dec r22 + brne 10b + movw r28, q0 + /* calculate W and store it in Q */ + ldi r19, 5 +30: + ldi r18, 16 + /* load initial index */ + + /* load values from hacktable */ + ldi r30, lo8(f0_hacktable-3) + ldi r31, hi8(f0_hacktable-3) + mov r16, r19 + lsl r16 + add r16, r19 + add r30, r16 + adc r31, r1 + lpm r21, Z+ + lpm r20, Z+ + lpm r16, Z+ +40: + ;call add_hx_to_w +add_hx_to_w: + movw r26, h0 + add r26, r16 + adc r27, r1 + rcall load32_from_Y + sbiw r28, 4 + lsl r20 + rol r21 + brcs 300f + /* addition */ + rcall add_X_to_32 + rjmp 500f +300: /* substract */ + rcall load_acc_from_X + sub r22, acc0 + sbc r23, acc1 + sbc r24, acc2 + sbc r25, acc3 + +500: + rcall store32_to_Y + subi r16, -4 + andi r16, 0x0f<<2 + dec r18 + brne 40b + movw r28, q0 + dec r19 + brne 30b + movw r26, h0 + /* xor m into h */ +; ldi r20, 64 + movw r26, h0 + movw r30, m0 + rcall memxor_short + sbiw r26, 60 +;--- + clr r17 + ldi r21, 15 + mov r8, r21 +50: + rcall load32_from_Y + sbiw r28, 4 + mov r20, r17 + rcall sn + inc r17 + cpi r17, 5 + brne 52f + clr r17 +52: + rcall add_X_to_32 + rcall store32_to_Y + + dec r8 + brne 50b +;--- + rcall load32_from_Y + clr r20 + rcall sn + movw r26, h0 + rcall add_X_to_32 + sbiw r26, 4 + sbiw r28, 4 + rcall store32_to_Y + sbiw r28, 4 + sbiw r28, 15*4 + movw r20, h0 + movw r22, m0 + + /* call f1*/ + movw r2, r28 f1: - movw r2, r24 movw r4, r22 movw r6, r20 movw r26, r2 -; movw r22, r4 -; movw r20, r6 clr r24 rcall expand1 movw r26, r2 @@ -634,20 +670,14 @@ f1: movw r24, r2 movw r22, r4 movw r20, r6 - ret -/******************************************************************************/ -/* - param q: r24:r25 - param m: r22:r23 - param h: r20:r21 -*/ -f2_1_shift_table: - .byte 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 - .byte 0x2B, 0x64, 0x66, 0x03, 0x51, 0x55, 0x87, 0x55 -f2_2_shift_table: - .byte (2<<1), (7<<1), (4<<1), (3<<1), (4<<1)+1, (6<<1)+1, (6<<1) - .byte 0 ; just for alignment + + /* call f2 */ +; pop_range 20, 25 +; push_range 20, 25 +; rcall printQ +; push r20 +; push r21 acc2 = 8 acc3 = 9 acc0 = 14 @@ -670,35 +700,23 @@ f2: adiw r26, 63 adiw r26, 1 movw q16_0, r26 - clr xl0 - clr xl1 - clr xl2 - clr xl3 - ldi r17, 8 -10: ld r0, X+ - eor xl0, r0 - ld r0, X+ - eor xl1, r0 - ld r0, X+ - eor xl2, r0 - ld r0, X+ - eor xl3, r0 - dec r17 - brne 10b -;--- /* calc XH */ - movw xh0, xl0 - movw xh2, xl2 - ldi r17, 8 -10: ld r0, X+ - eor xh0, r0 - ld r0, X+ - eor xh1, r0 - ld r0, X+ - eor xh2, r0 - ld r0, X+ - eor xh3, r0 + movw h0, r20 + movw r28, r22 + rcall load32_from_X + movw acc0, r22 + movw acc2, r24 + ldi r17, 15 +10: rcall load32_from_X + rcall eor32_to_acc + cpi r17, 9 + brne 15f + movw xl0, acc0 + movw xl2, acc2 +15: dec r17 brne 10b + movw xh0, acc0 + movw xh2, acc2 ;--- DBG ; push_range 22, 25 ; movw r22, xl0 @@ -711,8 +729,6 @@ f2: ;--- END DBG ;--- /* calc first half of h0..h15 */ - movw h0, r20 - movw r28, r22 movw r26, q16_0 ldi r17, 16 10: @@ -721,10 +737,14 @@ f2: ld acc2, Y+ ld acc3, Y+ ;--- - ldi r30, lo8(f2_1_shift_table-1) - ldi r31, hi8(f2_1_shift_table-1) movw r22, xh0 movw r24, xh2 + cpi r17, 9 + brge 15f + clr r1 + rjmp 26f +15: ldi r30, lo8(f2_1_shift_table-9) + ldi r31, hi8(f2_1_shift_table-9) add r30, r17 adc r31, r1 lpm r20, Z @@ -829,94 +849,23 @@ f2: rcall add_acc_to_Z inc r18 cpi r17, 5 - breq 20f - dec r17 + brne 20f + sbiw r26, 8*4 +20: dec r17 brne 10b - ret -20: sbiw r26, 8*4 - dec r17 - rjmp 10b - -/******************************************************************************/ -/* - param ctx: r24:r25 - param msg: r22:r23 -*/ -/* f0 - param q: r28:r29 (Y) - param h: r26:r27 (X) - param m: r30:r31 (Z) -*/ -/* f1 - param q: r24:r25 - param m: r22:r23 - param h: r20:r21 -*/ -/* f2 - param q: r24:r25 - param m: r22:r23 - param h: r20:r21 -*/ -.global bmw_small_nextBlock -.global bmw224_nextBlock -.global bmw256_nextBlock -bmw_small_nextBlock: -bmw224_nextBlock: -bmw256_nextBlock: - push_range 28, 29 - push_range 2, 17 - stack_alloc_large 32*4, r28, r29 - adiw r28, 1 -; push_range 28, 29 /* push Q */ -; push_range 22, 25 /* push M & H */ - /* increment counter */ - movw r26, r24 - movw r2, r26 - adiw r26, 63 - adiw r26, 1 - rcall load_acc_from_X - ldi r19, 1 - add acc0, r19 - adc acc1, r1 - adc acc2, r1 - adc acc3, r1 - st -X, acc3 - st -X, acc2 - st -X, acc1 - st -X, acc0 - /* call f0 */ - movw r30, r22 - movw r26, r24 - rcall f0 - /* call f1*/ - movw r24, r28 -; rcall printQ - rcall f1 - /* call f2 */ -; pop_range 20, 25 -; push_range 20, 25 -; rcall printQ -; push r20 -; push r21 - call f2 ;--- DBG ; pop r25 ; pop r24 ; ldi r22, 'H' ; rcall printX ;--- END DBG - stack_free_large3 32*4 + stack_free_large3 32*4+4 pop_range 2, 17 pop_range 28, 29 ret /******************************************************************************/ -/* - param ctx: r24:r25 - param msg: r22:r23 - param len: r20:r21 -*/ ctx0 = 2 ctx1 = 3 blc0 = 4 @@ -926,6 +875,22 @@ len1 = 29 buf0 = 6 buf1 = 7 +load32_from_Z_stub: + movw r30, ctx0 + adiw r30, 60 + ldd r21, Z+4 + ldd r22, Z+5 + ldd r23, Z+6 + ldd r24, Z+7 + ret + +/******************************************************************************/ +/* + param ctx: r24:r25 + param msg: r22:r23 + param len: r20:r21 +*/ + .global bmw_small_lastBlock .global bmw224_lastBlock .global bmw256_lastBlock @@ -1020,12 +985,7 @@ bmw256_lastBlock: st X+, r1 dec r20 brne 350b - movw r30, ctx0 - adiw r30, 60 - ldd r21, Z+4 - ldd r22, Z+5 - ldd r23, Z+6 - ldd r24, Z+7 + rcall load32_from_Z_stub subi r21, 1 sbc r22, r1 sbc r23, r1 @@ -1035,12 +995,7 @@ bmw256_lastBlock: bmw_small_nextBlock(ctx, pctx.buffer); */ 400: - movw r30, ctx0 - adiw r30, 60 - ldd r21, Z+4 - ldd r22, Z+5 - ldd r23, Z+6 - ldd r24, Z+7 + rcall load32_from_Z_stub 410: clr r25 lsl r21 @@ -1072,16 +1027,15 @@ bmw256_lastBlock: pctx.buffer[i*4] = i+0xa0; } */ - ldi r18, 0xa0 - ldi r19, 0xaa + ldi r22, 0xa0 + ldi r23, 0xaa + ldi r24, 0xaa + ldi r25, 0xaa movw r26, buf0 500: - st X+, r18 - st X+, r19 - st X+, r19 - st X+, r19 - inc r18 - sbrs r18, 4 + rcall store32_to_X + inc r22 + sbrs r22, 4 rjmp 500b /* bmw_small_nextBlock((bmw_small_ctx_t*)&pctx, ctx->h); memcpy(ctx->h, pctx.buffer, 64); @@ -1195,19 +1149,20 @@ ctx0 = 2 ctx1 = 3 msg0 = 4 msg1 = 5 -len0 = 6 -len1 = 7 +len0 = 28 +len1 = 29 len2 = 8 len3 = 9 -dst0 = 10 -dst1 = 11 +dst0 = 6 +dst1 = 7 .global bmw224 bmw224: push r16 clr r16 bmw_small_all: - push_range 2, 11 + push_range 2, 9 + push_range 28, 29 stack_alloc_large 64+4 adiw r30, 1 movw ctx0, r30 @@ -1228,8 +1183,7 @@ bmw_small_all: movw r24, ctx0 movw r22, msg0 rcall bmw_small_nextBlock - ldi r20, 2 - sub len1, r20 + subi len1, 2 sbc len2, r1 sbc len3, r1 ldi r20, 64 @@ -1249,7 +1203,8 @@ bmw_small_all: adc r31, r1 icall stack_free_large 64+4 - pop_range 2, 11 + pop_range 28, 29 + pop_range 2, 9 pop r16 ret @@ -1280,10 +1235,7 @@ bmw224_init: ldi r24, 0x01 ldi r25, 0x00 bmw_small_init: - st X+, r22 - st X+, r23 - st X+, r24 - st X+, r25 + rcall store32_to_X ldi r18, 16-1 ldi r20, 0x04 1: @@ -1291,10 +1243,7 @@ bmw_small_init: adc r23, r20 adc r24, r20 adc r25, r20 - st X+, r22 - st X+, r23 - st X+, r24 - st X+, r25 + rcall store32_to_X dec r18 brne 1b st X+, r1