noekeon optimized++

[avr-crypto-lib.git] / noekeon_asm.S

diff --git a/noekeon_asm.S b/noekeon_asm.S
index 8ebcc93c6578ce3b13698cf4e9bff2cf084c7278..146cbb4a4d1f3f11a0f4fd1affc86a84783f888d 100644 (file)
--- a/noekeon_asm.S
+++ b/noekeon_asm.S
@@ -68,6 +68,18 @@
        clr r1
 .endm
 
+push_all_func:
+       pop r31
+       pop r30
+       push_all
+       ijmp
+
+pop_all_func:
+    pop r31
+    pop r30
+    pop_all
+    ijmp
+
 .macro xchg a b
        eor \a, \b
        eor \b, \a
@@ -275,6 +287,7 @@ theta:
        ret
 
 /******************************************************************************/
+#ifndef NOEKEON_NO_ENC
 ; === noekeon_enc ===
 ;
 ;  param1: pointer to buffer/state (r24,r25)
@@ -282,7 +295,7 @@ theta:
 ;
 .global noekeon_enc
 noekeon_enc:
-       push_all
+       rcall push_all_func
        /* load state */
        movw r26, r22
        ldi r28, 2
@@ -312,7 +325,7 @@ noekeon_enc:
        lpm r0, Z
        push r0
 3:
-       call round /* pops rc2 & rc1 */
+       rcall round /* pops rc2 & rc1 */
        pop r22
        dec r22
        push r22
@@ -322,7 +335,7 @@ noekeon_enc:
 
        ldi r22, 0xD4
        eor state0_3, r22
-       call theta
+       rcall theta
 
        pop r31
        pop r30
@@ -335,10 +348,13 @@ noekeon_enc:
        dec r22
        brne 1b
        
-       pop_all
+       rcall pop_all_func
        ret
+#endif
 /******************************************************************************/
 /******************************************************************************/
+#ifndef NOEKEON_NO_DEC
+
 ; === noekeon_dec ===
 ;
 ;  param1: pointer to buffer/state (r24,r25)
@@ -346,7 +362,7 @@ noekeon_enc:
 ;
 .global noekeon_dec
 noekeon_dec:
-       push_all
+       rcall push_all_func
        /* allocate 16 bytes on the stack */
        in r30, _SFR_IO_ADDR(SPL)
        in r31, _SFR_IO_ADDR(SPH)
@@ -377,7 +393,7 @@ noekeon_dec:
 
        movw r26, r30
        sbiw r26, 16 /* set X back to begining of stack key */
-       call theta
+       rcall theta
        
        /* mov state to stackkey */
        clr r29
@@ -419,7 +435,7 @@ noekeon_dec:
        push r0
        push r1
 3:
-       call round /* pops rc2 & rc1 */
+       rcall round /* pops rc2 & rc1 */
        pop r22
        dec r22
        push r22
@@ -427,7 +443,7 @@ noekeon_dec:
 ;----
        pop r22
 
-       call theta
+       rcall theta
        ldi r22, 0x80
        eor state0_3, r22
        
@@ -450,16 +466,18 @@ write_state_back:
        adiw r30, 16 
        out  _SFR_IO_ADDR(SPH), r31
        out  _SFR_IO_ADDR(SPL), r30
-       pop_all
+       rcall pop_all_func
        ret
+#endif
 /******************************************************************************/
        
+       
 round: 
        pop r24
        pop r25
        pop r1
        eor state0_3, r1
-       call theta
+       rcall theta
        pop r1
        eor state0_3, r1
        push r25
@@ -467,9 +485,9 @@ round:
 pi_gamma_pi:
        ldi r30, pm_lo8(bigendian_rotl32)
        ldi r31, pm_hi8(bigendian_rotl32)
-       call pi
+       rcall pi
        /* pi1 done; now gamma */
-       call gamma_1
+       rcall gamma_1
        /* a[0] <-> a[3] */
        xchg state0_0, state3_0
        xchg state0_1, state3_1
@@ -480,10 +498,10 @@ pi_gamma_pi:
        op32 eor, state2, state1
        op32 eor, state2, state3
 
-       call gamma_1
+       rcall gamma_1
        ldi r30, pm_lo8(bigendian_rotr32)
        ldi r31, pm_hi8(bigendian_rotr32)
-       call pi
+       rcall pi
        ret
        
 gamma_1:
@@ -573,6 +591,9 @@ void noekeon_init(void* key, noekeon_ctx_t* ctx){
        noekeon_enc(ctx, nullv);
 }
 */
+
+#ifndef NOEKEON_NO_INIT
+
 .global noekeon_init
 noekeon_init:
 ; === noekeon_init ===
@@ -606,7 +627,7 @@ noekeon_init:
        
        sbiw r26, 16
        movw r24, r26
-       call noekeon_enc
+       rcall noekeon_enc
        
        in r30, _SFR_IO_ADDR(SPL)
        in r31, _SFR_IO_ADDR(SPH)
@@ -615,6 +636,6 @@ noekeon_init:
        out  _SFR_IO_ADDR(SPL), r30     
        ret
        
-
+#endif