Free 15 bytes by reorganizing code (1023 to 1008 bytes)

* Load the last nibble of the CHIP-8 instruction (PIR+1) in the fetch routine
* Move some branches to allow fallthrough

CHIPOS68.asm

@@ -6,6 +6,7 @@
 ;     DREAM-6800 OPERATING SYSTEM WITH CHIP8 LANGUAGE INTERPRETER
 ;
 ;       ORIGINATED BY MICHAEL J BAUER, DEAKIN UNIVERSITY, 1978
+;    MODIFICATIONS BY TOBIAS V LANGHOFF, UNIVERSITY OF OSLO, 2020
 ;
 ;                  www.mjbauer.biz/DREAM6800.htm
 ;
@@ -81,45 +82,46 @@ FETCH:  LDX     PPC         ; POINT TO NEXT INSTR
         STX     PIR
         STX     ZHI         ; SAVE ADRS (MMM)
         JSR     SKIP2       ; BUMP PRGM-CTR
-        LDAA    ZHI         ; MASK OFF ADRS
-        ANDA    #$0F
-        STAA    ZHI
+        LDAB    ZHI         ; MASK OFF ADRS
+        ANDB    #$0F
+        STAB    ZHI
         BSR     FINDV       ; EXTRACT VX ALSO
-        STAA    VX          ; STASH VX
+        STAB    VX          ; STASH VX
         STX     VXLOC       ; SAVE LOCATION OF VX
-        LDAA    PIR+1       ; FIND Y
-        LSRA
-        LSRA
-        LSRA
-        LSRA
+        LDAB    PIR+1       ; FIND Y
+        TBA                 ; STASH KK
+        LSRB
+        LSRB
+        LSRB
+        LSRB
         BSR     FINDV       ; EXTRACT VY
-        STAA    VY          ; STASH VY
+        STAB    VY          ; STASH VY
 EXEC:   LDX     #JUMTAB-2   ; POINT TO JUMP TABLE
-        LDAA    PIR         ; EXTRACT MSD
-        ANDA    #$F0
+        LDAB    PIR         ; EXTRACT MSD
+        ANDB    #$F0
 EXEl:   INX                 ; FIND ROUTINE ADRS
         INX
-        SUBA    #$10
+        SUBB    #$10
         BCC     EXEl        ; BRANCH IF HIGHER OR SAME
         LDX     0,X         ; LOAD ROUTINE ADRS
         JSR     0,X         ; PERFORM ROUTINE
         BRA     FETCH       ; NEXT INSTR...
 FINDV:  LDX     #VO-1       ; POINT TO VARIABLES TABLE
 FIND1:  INX                 ; FIND LOCN VX
-        DECA
+        DECB
         BPL     FIND1
-        LDAA    0,X         ; FETCH VX FROM TABLE
+        LDAB    0,X         ; FETCH VX FROM TABLE
         RTS
 ;
 ; JUMP TABLE(ROUTINE ADDRESSES)
 ;
 JUMTAB: FDB     EXCALL      ; ERASE, RET, CALL, NOP
         FDB     GOTO        ; GOTO MMM
         FDB     DOSUB       ; DO MMM
-        FDB     SKFEK       ; SKF VX=KK
-        FDB     SKFNK       ; SKF VX#KK
+        FDB     SKFEQ       ; SKF VX=KK
+        FDB     SKFNE       ; SKF VX#KK
         FDB     SKFEV       ; SKF VX=VY
-        FDB     LETK        ; Vx=KK
+        FDB     PUTVX       ; Vx=KK
         FDB     LETVK       ; VX=VX+KK
         FDB     LETVV       ; VX=[VX][+-&!]VY
         FDB     SKFNV       ; SKF  VX#VY
@@ -134,12 +136,12 @@ JUMTAB: FDB     EXCALL      ; ERASE, RET, CALL, NOP
 ;
 EXCALL: LDAB    PIR         ; GET INSTR REG
         BNE     CALL
-        LDAA    PIR+1
-        CMPA    #$E0
-        BEQ     ERASE
         CMPA    #$EE
         BEQ     RETDO
-        RTS                 ; NOP, FETCH
+        CMPA    #$E0
+        BNE     RETMON      ; NOP, FETCH
+        NOP                 ; PADDING FOR FALLTHROUGH
+        NOP                 ; PADDING FOR FALLTHROUGH
 ;
         ORG     $C079
 ;
@@ -152,7 +154,7 @@ FILL:   STAA    0,X         ; FILL SCREEN WITH ACC-A
         INX
         CPX     #ENDBUF     ; DONE?
         BNE     FILL
-        RTS
+RETMON: RTS
 RETDO:  TSX                 ; SAVE REAL SP
         LDS     PSP
         PULA
@@ -188,14 +190,9 @@ DOSUB:  TSX                 ; SAVE SP
 ;
 ; CONDITIONAL SKIP ROUTINES
 ;
-SKFEK:  LDAA    PIR+1       ; GET KK
 SKFEQ:  CMPA    VX
         BEQ     SKIP2
         RTS
-SKFNK:  LDAA    PIR+1       ; GET KK
-SKFNE:  CMPA    VX
-        BNE     SKIP2
-        RTS
 SKFEV:  LDAA    VY          ; GET VY
         BRA     SKFEQ
 SKFNV:  LDAA    VY
@@ -212,23 +209,23 @@ SKFKEY: JSR     KEYINP      ; INTERROGATE KEYBOARD
         CMPB    PIR+1       ; SKF VX#KEY
         BEQ     SKIP2
         RTS                 ; NO KEY GO FETCH
-SKFK1:  LDAB    #$9E
-        CMPB    PIR+1       ; WHAT INSTRN?
+SKFK1:  CMPA    #$9E
         BEQ     SKFEQ
-        BRA     SKFNE
+SKFNE:  CMPA    VX
+        BNE     SKIP2
+        RTS
 ;
 ; ARITHMETIC/LOGIC ROUTINES
 ;
-LETK:   LDAA    PIR+1       ; GET KK
-        BRA     PUTVX
-LETVK:  LDAA    PIR+1
-        ADDA    VX
+LETVK:  ADDA    VX
         BRA     PUTVX
 RANDV:  BSR     RANDOM      ; GET RANDOM BYTE
         ANDA    PIR+1
         BRA     PUTVX
-LETVV:  LDAA    VX
-        LDAB    PIR+1
+
+LETVV:  TAB
+        LDAA    VX
+
         ANDB    #$0F        ; EXTRACT N
         BNE     LETV1
         LDAA    VY          ; VX=VY
@@ -409,9 +406,9 @@ MOVX1:  PULA                ; GET NEXT V
         RTS
 ;
 ; DISPLAY ROUTINES
-;
-SHOW:   LDAB    PIR+1       ; GET N (OPCODE LSB)
+SHOW:   TAB                 ; GET N (OPCODE LSB)
         CLR     VF          ; CLEAR OVERLAP FLAG
+        NOP                 ; PADDING FOR FALLTHROUGH
 ;
         ORG     $C224
 ;

README.md

@@ -19,4 +19,4 @@ p2bin -l 0 CHIPOS68.p
 
 The CHIPOS code in this repo is also available for download on [Michael J. Bauer's DREAM 6800 website](http://www.mjbauer.biz/DREAM6800.htm). Michael has stated to me in private correspondence that the code is in the public domain.
 
-This version contains several origins. These are not needed for assembly (besides the last one that makes sure the interrupt vectors are located at the highest addresses), but they also serve to denote sections of the code that programs will expect to stay put, should you wish to edit it while still staying CHIPOS-compatible. These addresses were all found in the document "CHIPOS SUBROUTINES (& Calling Sequences)" from Michael J. Bauer's website.
+This version is a few (18) bytes smaller than the original CHIPOS, to make room for extra functionality. Therefore, it contains several origins to remain compatible with the addresses found in the document "CHIPOS SUBROUTINES (& Calling Sequences)" from Michael J. Bauer's website.