[keccak-asm] removing unnecessary c and d fields from context

[avr-crypto-lib.git] / keccak / keccak-asm.S

diff --git a/keccak/keccak-asm.S b/keccak/keccak-asm.S
index 3b3a48818fdc84a47547b64d61e6b00ad1eca3a9..7b7c6cf3144bb4374831c4cb5dbb1ecb9f66d420 100644 (file)
--- a/keccak/keccak-asm.S
+++ b/keccak/keccak-asm.S
@@ -43,10 +43,6 @@ ctx_a:
        .struct ctx_a + 8 * 5 * 5
 ctx_r:
        .struct ctx_r + 2
-ctx_c:
-       .struct ctx_c + 2
-ctx_d:
-       .struct ctx_d + 1
 ctx_bs:
 
        .section .text
@@ -329,7 +325,7 @@ chi_step:
 keccak_nextBlock:
        movw ZL, r24
        subi ZL, lo8(-ctx_bs)
-       sbci ZL, hi8(-ctx_bs)
+       sbci ZH, hi8(-ctx_bs)
        ld r20, Z
        movw XL, r24
        movw ZL, r22
@@ -340,8 +336,10 @@ keccak_nextBlock:
        st X+, r22
        dec r20
        brne 10b
+       .endfunc
 
        .global keccak_f1600
+       .func keccak_f1600
 keccak_f1600:
        push_range 2, 9
        push r16
@@ -542,15 +540,17 @@ icall_r16_times:
 */
        ; 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
+       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
 
@@ -619,5 +619,73 @@ icall_r16_times:
        pop_range 28, 29
        pop r16
        pop_range 2, 9
+       ret
+       .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, r24
+       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
+       rjmp 40f
+       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
+