further reduction of BMW tiny size (1888 bytes)

diff --git a/bmw/bmw_small-tinyasm.S b/bmw/bmw_small-tinyasm.S
index ddad3c9f68fdf9cadf96a36a08a8f9ce0e40b3ef..de7ff5a2f62360952f9700b0d7fdc3020cf5f7d0 100644 (file)
--- a/bmw/bmw_small-tinyasm.S
+++ b/bmw/bmw_small-tinyasm.S
@@ -189,31 +189,6 @@ h1 = 5
 m0 = 6
 m1 = 7
 
-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 30f
-       /* addition */
-       rcall add_X_to_32
-       rjmp store32_to_Y;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:
-       rjmp store32_to_Y
-;      rcall store32_to_Y
-;      ret
 
 /******************************************************************************/
 load32_from_X:
@@ -269,112 +244,6 @@ f0_s_table:
        .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:
-       rcall load32_from_Y
-       sbiw r28, 4
-       lpm r20, Z+
-       movw r2, r30
-       rcall sn
-       movw r30, r2
-
-       rcall add_X_to_32
-       rcall store32_to_Y
-
-       dec r8
-       brne 50b
-;---
-       rcall load32_from_Y
-       clr r20
-       rcall sn
-       movw r30, r2
-       movw r26, h0
-       rcall add_X_to_32
-       sbiw r26, 4
-       st -Y, r25
-       st -Y, r24
-       st -Y, r23
-       st -Y, r22
-       sbiw r28, 15*4
-       movw r20, h0
-       movw r22, m0
-       ret
 
 /******************************************************************************/
 
@@ -538,13 +407,8 @@ expand1:
        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
+
 
 /******************************************************************************/
 /*
@@ -570,16 +434,20 @@ expand2:
        rcall add32_to_acc
        dec r19
        brne 10b
-       rcall load32_from_X
        ldi r20, 4
+       rcall load32_from_X
        rcall sn
        rcall add32_to_acc
-       rcall load32_from_X
        ldi r20, 5
+       rcall load32_from_X
        rcall sn
        rcall add32_to_acc
-
-       rjmp expand1_exit
+expand2_exit:
+       st X+, acc0
+       st X+, acc1
+       st X+, acc2
+       st X+, acc3
+       ret
 
 /******************************************************************************/
 /*
@@ -593,13 +461,193 @@ expand2:
   param h: r20:r21
   param j: r24
 */
+
+/******************************************************************************/
+/*
+  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
+
+/******************************************************************************/
+/*
+  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
+       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
+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 */
+       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
+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 */
+       ld r0, X+
+       sub r22, r0
+       ld r0, X+
+       sbc r23, r0
+       ld r0, X+
+       sbc r24, r0
+       ld r0, X+
+       sbc r25, r0
+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
+;---
+       ldi r30, lo8(f0_s_table)
+       ldi r31, hi8(f0_s_table)
+       ldi r21, 15
+       mov r8, r21
+50:
+       rcall load32_from_Y
+       sbiw r28, 4
+       lpm r20, Z+
+       movw r2, r30
+       rcall sn
+       movw r30, r2
+
+       rcall add_X_to_32
+       rcall store32_to_Y
+
+       dec r8
+       brne 50b
+;---
+       rcall load32_from_Y
+       clr r20
+       rcall sn
+       movw r30, r2
+       movw r26, h0
+       rcall add_X_to_32
+       sbiw r26, 4
+       st -Y, r25
+       st -Y, r24
+       st -Y, r23
+       st -Y, r22
+       sbiw r28, 15*4
+       movw r20, h0
+       movw r22, m0
+
+       /* call f1*/
+       movw r24, 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
@@ -619,20 +667,14 @@ 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
@@ -800,77 +842,11 @@ 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