#include "avr-asm-macros.S"
shiftcodetable:
- .byte 0x00 ; 0
+; .byte 0x00 ; 0
+shiftcodetable_1:
.byte 0x01 ; 1
.byte 0x02 ; 2
.byte 0x03 ; 3
.byte 0x2A ; 14
.byte 0x29 ; 15
.byte 0x20 ; 16
- .byte 0x21 ; 17 unused but necesseray for padding
+; .byte 0x21 ; 17 unused but necesseray for padding
.global rotl_addel
rotl_addel:
andi r20, 0x0f
- inc r20
- ldi r30, lo8(shiftcodetable)
- ldi r31, hi8(shiftcodetable)
+ ldi r30, lo8(shiftcodetable_1)
+ ldi r31, hi8(shiftcodetable_1)
add r30, r20
adc r31, r1
lpm r20, Z
movw r22, r30
2: bst r20, 3
andi r20, 0x07
- brne 3f
- ret
+ breq some_ret
3:
brts rotr32; 4f
rjmp rotl32
.long 0x9555554c, 0x9aaaaaa1, 0x9ffffff6, 0xa555554b
/*******************************************************************************
-* uint32_t addelment(uint8_t j, const uint32_t* m, const uint32_t* h){
+* uint32_t addelment(uint8_t j, const uint32_t *m, const uint32_t *h){
* uint32_t r;
* r = pgm_read_dword(k_lut+j);
* r += rotl_addel(((uint32_t*)m)[j&0xf], j+0);
mov j, r24
movw h0, r20
movw m0, r22
- mov r25, r24
- lsl r25
- lsl r25
+ lsl r24
+ lsl r24
ldi r30, lo8(const_lut)
ldi r31, hi8(const_lut)
- add r30, r25
+ add r30, r24
adc r31, r1
lpm acc0, Z+
lpm acc1, Z+
ret
/*******************************************************************************
-* uint32_t bmw_small_expand1(uint8_t j, const void* m, const void* h, const uint32_t* q){
+* uint32_t bmw_small_expand1(uint8_t j, const void *m, const void *h, const uint32_t *q){
* uint32_t(*s[])(uint32_t) = {bmw_small_s1, bmw_small_s2, bmw_small_s3, bmw_small_s0};
* uint32_t r;
* uint8_t i;
ret
/*******************************************************************************
-* uint32_t bmw_small_expand2(uint8_t j, const void* m, const void* h, const uint32_t* q){
+* uint32_t bmw_small_expand2(uint8_t j, const void *m, const void *h, const uint32_t *q){
* uint32_t(*rf[])(uint32_t) = {bmw_small_r1, bmw_small_r2, bmw_small_r3,
* bmw_small_r4, bmw_small_r5, bmw_small_r6,
* bmw_small_r7};
rjmp expand1_exit
/*******************************************************************************
-* void bmw_small_f1(uint32_t* q, const void* m, const void* h){
+* void bmw_small_f1(uint32_t *q, const void *m, const void *h){
* uint8_t i;
* q[16] = bmw_small_expand1(0, m, h, q);
* q[17] = bmw_small_expand1(1, m, h, q);
* uint16_t hack_table[5] PROGMEM = { 0x0311, 0xDDB3, 0x2A79, 0x07AA, 0x51C2 };
* uint8_t offset_table[5] PROGMEM = { 4+16, 6+16, 9+16, 12+16, 13+16 };
*
-* void bmw_small_f0(uint32_t* h, const void* m, uint32_t* q){
+* void bmw_small_f0(uint32_t *h, const void *m, uint32_t *q){
* uint16_t hack_reg;
* uint8_t c,i,j;
* uint32_t(*s[])(uint32_t)={ bmw_small_s0, bmw_small_s1, bmw_small_s2,
ret
/*******************************************************************************
-* void bmw_small_f2(uint32_t* h, const uint32_t* q, const void* m){
+* void bmw_small_f2(uint32_t *h, const uint32_t *q, const void *m){
* uint32_t xl=0, xh;
* uint8_t i;
* for(i=16;i<24;++i){
; pop_range 28, 29
ret
+#if DEBUG_FUNCTIONS
+
cli_putb:
push r2
push_range 18, 26
pop_range 18, 31
ret
+#endif
+
/*******************************************************************************
-* void bmw_small_nextBlock(bmw_small_ctx_t* ctx, const void* block){
+* void bmw_small_nextBlock(bmw_small_ctx_t *ctx, const void *block){
* uint32_t q[32];
* dump_x(block, 16, 'M');
* bmw_small_f0(ctx->h, block, q);
q0 = 6
q1 = 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, 30, 31
ret
+/*******************************************************************************
+* void bmw224_init(bmw224_ctx_t *ctx){
+* uint8_t i;
+* ctx->h[0] = 0x00010203;
+* for(i=1; i<16; ++i){
+* ctx->h[i] = ctx->h[i-1]+ 0x04040404;
+* }
+* ctx->counter=0;
+* }
+*
+* param ctx: r24:r25
+*/
+.global bmw224_init
+bmw224_init:
+ movw r26, r24
+ ldi r22, 0x03
+ ldi r23, 0x02
+ ldi r24, 0x01
+ ldi r25, 0x00
+bmw_small_init:
+ st X+, r22
+ st X+, r23
+ st X+, r24
+ st X+, r25
+ ldi r18, 16-1
+ ldi r20, 0x04
+1:
+ add r22, r20
+ adc r23, r20
+ adc r24, r20
+ adc r25, r20
+ st X+, r22
+ st X+, r23
+ st X+, r24
+ st X+, r25
+ dec r18
+ brne 1b
+ st X+, r1
+ st X+, r1
+ st X+, r1
+ st X+, r1
+ ret
+
+.global bmw256_init
+bmw256_init:
+ movw r26, r24
+ ldi r22, 0x43
+ ldi r23, 0x42
+ ldi r24, 0x41
+ ldi r25, 0x40
+ rjmp bmw_small_init
+
+/*******************************************************************************
+* void bmw_small_lastBlock(bmw_small_ctx_t *ctx, const void *block, uint16_t length_b){
+* struct {
+* uint8_t buffer[64];
+* uint32_t ctr;
+* } pctx;
+* while(length_b >= BMW_SMALL_BLOCKSIZE){
+* bmw_small_nextBlock(ctx, block);
+* length_b -= BMW_SMALL_BLOCKSIZE;
+* block = (uint8_t*)block + BMW_SMALL_BLOCKSIZE_B;
+* }
+* memset(pctx.buffer, 0, 64);
+* memcpy(pctx.buffer, block, (length_b+7)/8);
+* pctx.buffer[length_b>>3] |= 0x80 >> (length_b&0x07);
+* if(length_b+1>64*8-64){
+* bmw_small_nextBlock(ctx, pctx.buffer);
+* memset(pctx.buffer, 0, 64-8);
+* ctx->counter -= 1;
+* }
+* *((uint64_t*)&(pctx.buffer[64-8])) = (uint64_t)(ctx->counter*512LL)+(uint64_t)length_b;
+* bmw_small_nextBlock(ctx, pctx.buffer);
+* uint8_t i;
+* memset(pctx.buffer, 0xaa, 64);
+* for(i=0; i<16;++i){
+* pctx.buffer[i*4] = i+0xa0;
+* }
+* bmw_small_nextBlock((bmw_small_ctx_t*)&pctx, ctx->h);
+* memcpy(ctx->h, pctx.buffer, 64);
+* }
+*
+* param ctx: r24:r25
+* param block: r22:r23
+* param length_b: r20:r21
+*/
+ctx0 = 2
+ctx1 = 3
+blc0 = 4
+blc1 = 5
+len0 = 28
+len1 = 29
+buf0 = 6
+buf1 = 7
+
+.global bmw_small_lastBlock
+.global bmw224_lastBlock
+.global bmw256_lastBlock
+bmw_small_lastBlock:
+bmw224_lastBlock:
+bmw256_lastBlock:
+/* while(length_b >= BMW_SMALL_BLOCKSIZE){
+ bmw_small_nextBlock(ctx, block);
+ length_b -= BMW_SMALL_BLOCKSIZE;
+ block = (uint8_t*)block + BMW_SMALL_BLOCKSIZE_B;
+ }
+*/
+ push_range 2, 7
+ push_range 28, 29
+ movw ctx0, r24
+ movw blc0, r22
+ movw len0, r20
+1:
+ cpi len1, hi8(512)
+ brlo 2f
+ movw r24, ctx0
+ movw r22, blc0
+ rcall bmw_small_nextBlock
+ ldi r24, 64
+ add blc0, r24
+ adc blc1, r1
+ subi len1, hi8(512)
+ rjmp 1b
+2:
+/* struct {
+ uint8_t buffer[64];
+ uint32_t ctr;
+ } pctx;
+*/
+ stack_alloc_large 68
+ adiw r30, 1
+ movw buf0, r30
+/* memset(pctx.buffer, 0, 64);
+ memcpy(pctx.buffer, block, (length_b+7)/8);
+ pctx.buffer[length_b>>3] |= 0x80 >> (length_b&0x07);
+*/ movw r24, len0
+ lsr r25
+ ror r24
+ lsr r24
+ lsr r24
+; inc r24
+ ldi r23, 63
+ sub r23, r24
+ movw r26, blc0
+ tst r24
+ breq 301f
+30: ld r20, X+
+ st Z+, r20
+ dec r24
+ brne 30b
+301:
+ clr r20
+ mov r21, len0
+ ldi r24, 0x80
+ andi r21, 0x07
+ breq 305f
+ ld r20, X+
+303:
+ lsr r24
+ dec r21
+ brne 303b
+305:
+ or r20, r24
+ st Z+, r20
+ tst r23
+ breq 32f
+31: st Z+, r1
+ dec r23
+ brne 31b
+32:
+/* if(length_b+1>64*8-64){ ; = 64*7-1 = 447 max(length_b)=511
+ bmw_small_nextBlock(ctx, pctx.buffer);
+ memset(pctx.buffer, 0, 64-8);
+ ctx->counter -= 1;
+ }
+*/
+ tst len1
+ breq 400f
+ cpi len0, 192
+ brlo 400f
+ movw r24, ctx0
+ movw r22, buf0
+ rcall bmw_small_nextBlock
+ movw r26, buf0
+ ldi r20, 64-8
+350:
+ 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
+ subi r21, 1
+ sbc r22, r1
+ sbc r23, r1
+ sbc r24, r1
+ rjmp 410f
+/* *((uint64_t*)&(pctx.buffer[64-8])) = (uint64_t)(ctx->counter*512LL)+(uint64_t)length_b;
+ 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
+410:
+ clr r25
+ lsl r21
+ rol r22
+ rol r23
+ rol r24
+ rol r25
+ mov r20, len0
+ add r21, len1
+ adc r22, r1
+ adc r23, r1
+ adc r24, r1
+ adc r25, r1
+ movw r30, buf0
+ adiw r30, 64-8
+ st Z+, r20
+ st Z+, r21
+ st Z+, r22
+ st Z+, r23
+ st Z+, r24
+ st Z+, r25
+ st Z+, r1
+ st Z+, r1
+ movw r24, ctx0
+ movw r22, buf0
+ rcall bmw_small_nextBlock
+/* memset(pctx.buffer, 0xaa, 64);
+ for(i=0; i<16;++i){
+ pctx.buffer[i*4] = i+0xa0;
+ }
+*/
+ ldi r18, 0xa0
+ ldi r19, 0xaa
+ movw r26, buf0
+500:
+ st X+, r18
+ st X+, r19
+ st X+, r19
+ st X+, r19
+ inc r18
+ sbrs r18, 4
+ rjmp 500b
+/* bmw_small_nextBlock((bmw_small_ctx_t*)&pctx, ctx->h);
+ memcpy(ctx->h, pctx.buffer, 64);
+*/
+ movw r24, buf0
+ movw r22, ctx0
+ rcall bmw_small_nextBlock
+ ldi r18, 64
+ movw r26, ctx0
+ movw r30, buf0
+600:
+ ld r20, Z+
+ st X+, r20
+ dec r18
+ brne 600b
+
+ stack_free_large 68
+ pop_range 28, 29
+ pop_range 2, 7
+ ret
+
+/*******************************************************************************
+* void bmw224_ctx2hash(void *dest, const bmw224_ctx_t *ctx){
+* memcpy(dest, &(ctx->h[9]), 224/8);
+* }
+*
+* param dest: r24:r25
+* param ctx: r22:r23
+*/
+.global bmw224_ctx2hash
+bmw224_ctx2hash:
+ movw r26, r24
+ movw r30, r22
+ adiw r30, 9*4
+ ldi r22, 28
+ rjmp 1f
+/*******************************************************************************
+* void bmw256_ctx2hash(void *dest, const bmw256_ctx_t *ctx){
+* memcpy(dest, &(ctx->h[8]), 256/8);
+* }
+*
+* param dest: r24:r25
+* param ctx: r22:r23
+*/
+.global bmw256_ctx2hash
+bmw256_ctx2hash:
+ movw r26, r24
+ movw r30, r22
+ adiw r30, 8*4
+ ldi r22, 32
+1:
+ ld r23, Z+
+ st X+, r23
+ dec r22
+ brne 1b
+ ret
+/*******************************************************************************
+* void bmw256(void *dest, const void *msg, uint32_t length_b){
+* bmw_small_ctx_t ctx;
+* bmw256_init(&ctx);
+* while(length_b>=BMW_SMALL_BLOCKSIZE){
+* bmw_small_nextBlock(&ctx, msg);
+* length_b -= BMW_SMALL_BLOCKSIZE;
+* msg = (uint8_t*)msg + BMW_SMALL_BLOCKSIZE_B;
+* }
+* bmw_small_lastBlock(&ctx, msg, length_b);
+* bmw256_ctx2hash(dest, &ctx);
+* }
+*
+* param dest: r24:r25
+* param msg: r22:r23
+* param length_b: r18:r21
+*/
+ctx0 = 2
+ctx1 = 3
+msg0 = 4
+msg1 = 5
+len0 = 6
+len1 = 7
+len2 = 8
+len3 = 9
+dst0 = 10
+dst1 = 11
+.global bmw256
+bmw256:
+ push r16
+ ldi r16, 1
+ rjmp bmw_small_all
+/*******************************************************************************
+* void bmw224(void *dest, const void *msg, uint32_t length_b){
+* bmw_small_ctx_t ctx;
+* bmw224_init(&ctx);
+* while(length_b>=BMW_SMALL_BLOCKSIZE){
+* bmw_small_nextBlock(&ctx, msg);
+* length_b -= BMW_SMALL_BLOCKSIZE;
+* msg = (uint8_t*)msg + BMW_SMALL_BLOCKSIZE_B;
+* }
+* bmw_small_lastBlock(&ctx, msg, length_b);
+* bmw224_ctx2hash(dest, &ctx);
+* }
+*
+* param dest: r24:r25
+* param msg: r22:r23
+* param length_b: r18:r21
+*/
+ctx0 = 2
+ctx1 = 3
+msg0 = 4
+msg1 = 5
+len0 = 6
+len1 = 7
+len2 = 8
+len3 = 9
+dst0 = 10
+dst1 = 11
+.global bmw224
+bmw224:
+ push r16
+ clr r16
+bmw_small_all:
+ push_range 2, 11
+ stack_alloc_large 64+4
+ adiw r30, 1
+ movw ctx0, r30
+ movw dst0, r24
+ movw msg0, r22
+ movw len0, r18
+ movw len2, r20
+ movw r24, ctx0
+ ldi r30, pm_lo8(init_lut)
+ ldi r31, pm_hi8(init_lut)
+ add r30, r16
+ adc r31, r1
+ icall
+20:
+ mov r18, len2
+ or r18, len3
+ breq 50f
+ movw r24, ctx0
+ movw r22, msg0
+ rcall bmw_small_nextBlock
+ ldi r20, 2
+ sub len1, r20
+ sbc len2, r1
+ sbc len3, r1
+ ldi r20, 64
+ add msg0, r20
+ adc msg1, r1
+ rjmp 20b
+50:
+ movw r24, ctx0
+ movw r22, msg0
+ movw r20, len0
+ rcall bmw_small_lastBlock
+ movw r24, dst0
+ movw r22, ctx0
+ ldi r30, pm_lo8(c2h_lut)
+ ldi r31, pm_hi8(c2h_lut)
+ add r30, r16
+ adc r31, r1
+ icall
+ stack_free_large 64+4
+ pop_range 2, 11
+ pop r16
+ ret
+init_lut:
+ rjmp bmw224_init
+ rjmp bmw256_init
+c2h_lut:
+ rjmp bmw224_ctx2hash
+ rjmp bmw256_ctx2hash