; extern void __asm AsmRenderWorld(register __a0 long,register __a1 long); ; extern void __asm AsmClearFrame(register __a0 long *,register __d0 long,register __d1 long); ; extern void __asm AsmChunky2Planar(register __a0 long,register __a1 long); XDEF _AsmRenderWorld XDEF _AsmClearFrame XDEF _AsmChunky2Planar ;-------------------------------------------------------------------- ;PARAMETERS ; ; a0 = xBUFFER ; a1 = fBUFFER ViewRight EQU 256 ViewBottom EQU 128 ViewCenterY EQU 64 ; ViewBottom/2 fBUFF_SIZE EQU 32768 half_fBUFF EQU fBUFF_SIZE/2 X EQUR d0 yinc EQUR d1 yrinc equr d2 HC EQUR d3 temp1 EQUR d4 endY EQUR d5 errY EQUR d6 VIEWR EQUR d7 xBUFF EQUR a0 fBUFF EQUR a1 fBUFF2 equr a2 LHC EQUR a3 BMP EQUR a4 LBMP EQUR a5 wall1 EQUR a3 wall2 EQUR a5 VIEWC EQUR a6 section renderworld,code _AsmRenderWorld: INIT: movem.l d2-d7/a2-a6,-(sp) moveq #0,X ; X <- ViewLeft move.l X,LHC move.l X,LBMP move.l #ViewRight,VIEWR movea.l #ViewCenterY,VIEWC ; --------------------------------------------------------------------- xLoop: move.l (xBUFF)+,HC ; Fetch height, column from xBUFFER movea.l (xBUFF)+,BMP ; Fetch bitmap from xBUFFER cmp.l LHC,HC ; Is Height = LastHeight & Column = LastColumn? bne.b ScaleColumn ; No cmp.l LBMP,BMP ; Is Bitmap = LastBitmap? bne.b ScaleColumn ; No ; ------------------------------- CopyColumn: move.l fBUFF,-(sp) ; make a backup of fBUFF adda.l X,fBUFF ; startY = fBUFF = X adda.l #half_fBUFF,fBUFF move.l fBUFF,fBUFF2 cmp.w VIEWC,HC ; Is height > ViewCenterY+1? ble.b copySmall ; no ; ------------------------------- copyBig: move.l VIEWC,temp1 bra.b PreCpy copySmall: move.l HC,temp1 PreCpy: subq.l #1,temp1 COPY: suba.l VIEWR,fBUFF2 move.b -1(fBUFF),(fBUFF) adda.l VIEWR,fBUFF move.b -1(fBUFF2),(fBUFF2) dbra temp1,COPY EndCpy: move.l (sp)+,fBUFF ; restore fBUFF bra.b NextX ; ------------------------------- ScaleColumn: ; if any future unexplained problems occur, try checking for HC > 0 tst.w HC beq.b NextX move.l fBUFF,-(sp) adda.l X,fBUFF ; startY = fBUFF = X add.l #half_fBUFF,fBUFF movea.l fBUFF,fBUFF2 move.l HC,temp1 ; Initialize the wall bitmap ptrs swap temp1 ext.l temp1 ; temp1 = column lsl.w #6,temp1 ; temp1 = column * 64 move.l BMP,wall1 adda.l temp1,wall1 ; bitmap+column*64 lea 31(wall1),wall1 move.l wall1,wall2 ; wall2 = bitmap+column*64+63-scrY addq.l #1,wall1 ; wall1 = bitmap+column*64+scrY moveq #32,yrinc ; do scaling division moveq #0,yinc cmp.w yrinc,HC bgt.b SetErr DoDiv: divs.w HC,yrinc ; move.w yrinc,yinc ; yinc = 32 / height swap yrinc ; yrinc = 32 % height SetErr: move.w HC,errY ; errY = -height neg.w errY cmp.w VIEWC,HC ; Is height > ViewCenterY+1 ? ble.b scaleSmall ; no scaleBig: move.l VIEWC,endY bra.b MoreScaleInit ; ------------------------------- NextX: addq.w #1,X cmp.w VIEWR,X bne.b xLoop FIN: movem.l (sp)+,d2-d7/a2-a6 rts ; ------------------------------- scaleSmall: move.l HC,endY ; ------------------------------- MoreScaleInit: subq.l #1,endY tst.w yinc beq.b SCALE2 ; ------------------------------- SCALE1: suba.l VIEWR,fBUFF2 move.b (wall1),(fBUFF) adda.l yinc,wall1 move.b (wall2),(fBUFF2) suba.l yinc,wall2 add.w yrinc,errY blt.b Check1ScaleFin sub.w HC,errY addq.l #1,wall1 subq.l #1,wall2 Check1ScaleFin: adda.l VIEWR,fBUFF dbra endY,SCALE1 move.l HC,LHC move.l BMP,LBMP move.l (sp)+,fBUFF bra.b NextX ; ------------------------------- SCALE2: suba.l VIEWR,fBUFF2 move.b (wall1),(fBUFF) move.b (wall2),(fBUFF2) add.w yrinc,errY blt.b Check2ScaleFin sub.w HC,errY addq.l #1,wall1 subq.l #1,wall2 Check2ScaleFin: adda.l VIEWR,fBUFF dbra endY,SCALE2 move.l (sp)+,fBUFF move.l HC,LHC move.l BMP,LBMP bra.b NextX ; --------------------------------------------------------------------- _AsmClearFrame move.l d2,-(sp) move.l #half_fBUFF/4-1,d2 ClearTop: move.l d0,(a0)+ dbra d2,ClearTop move.l #half_fBUFF/4-1,d2 ClearBottom: move.l d1,(a0)+ dbra d2,ClearBottom move.l (sp)+,d2 rts ; ---------------------------------------------------------------------------- ; peterm/chunky4.s ; Basically the same as peterm/chunky3.s, except use Chris Hames' idea ; of temporary FAST buffers to allow longword writes to CHIP RAM. ;----------------------------------------------------------------------------- ; chunky2planar: (new Motorola syntax) ; a0 -> chunky pixels ; a1 -> plane0 (assume other 7 planes are allocated contiguously) PlaneStartOffset equ 404 ; bytes PlaneSize equ 8000 ; bytes PlaneLineOffset equ 8 ; bytes ViewBottom equ 128 LongsPerLine equ 8 fbuffer equr a0 plane0 equr a1 plane1 equr a2 plane2 equr a3 plane3 equr a4 section chunks,code _AsmChunky2Planar: movem.l d2-d7/a2-a4,-(sp) ; set up register constants ; move.l #$0f0f0f0f,d5 ; d5 = constant $0f0f0f0f ; previous not used for 4 bit only colours move.l #$55555555,d6 ; d6 = constant $55555555 move.l #$3333cccc,d7 ; d7 = constant $3333cccc lea PlaneStartOffset(plane0),plane0 move.w #PlaneSize,d5 movea.l plane0,plane1 adda.l d5,plane1 movea.l plane1,plane2 adda.l d5,plane2 movea.l plane2,plane3 adda.l d5,plane3 move.w #ViewBottom,d5 ; main loop (starts here) processes 8 chunky pixels at a time PreMain: moveq #32,d0 mainloop: ; d2 = a7a6a5a4a3a2a1a0 b7b6b5b4b3b2b1b0 c7c6c5c4c3c2c1c0 d7d6d5d4d3d2d1d0 move.l (a0)+,d2 ; 12 get next 4 chunky pixels in d0 ; d3 = e7e6e5e4e3e2e1e0 f7f6f5f4f3f2f1f0 g7g6g5g4g3g2g1g0 h7h6h5h4h3h2h1h0 move.l (a0)+,d3 ; 12 get next 4 chunky pixels in d1 ; d2 = d2 & 0f0f0f0f ; d2 = ........a3a2a1a0 ........b3b2b1b0 ........c3c2c1c0 ........d3d2d1d0 ; and.l d5,d2 ; 8 d5=$0f0f0f0f ; previous line not necessary for 4 bit only graphics ; d3 = d3 & 0f0f0f0f ; d3 = ........e3e2e1e0 ........f3f2f1f0 ........g3g2g1g0 ........h3h2h1h0 ; and.l d5,d3 ; 8 d5=$0f0f0f0f ; previous line not necessary for 4 bit only graphics ; d2 = (d2 << 4) | d3 ; d2 = a3a2a1a0e3e2e1e0 b3b2b1b0f3f2f1f0 c3c2c1c0g3g2g1g0 d3d2d1d0h3h2h1h0 lsl.l #4,d2 ; 16 or.l d3,d2 ; 8 ; d3 = ((d2 & 33330000) << 2) | (swap(d2) & 3333cccc) | ((d2 & 0000cccc) >> 2) ; d3 = a1a0c1c0e1e0g1g0 b1b0d1d0f1f0h1h0 a3a2c3c2e3e2g3g2 b3b2d3d2f3f2h3h2 move.l d2,d3 ; 4 and.l d7,d3 ; 8 d7=$3333cccc move.w d3,d1 ; 4 clr.w d3 ; 4 lsl.l #2,d3 ; 12 lsr.w #2,d1 ; 10 or.w d1,d3 ; 4 swap d2 ; 4 and.l d7,d2 ; 8 d7=$3333cccc or.l d2,d3 ; 8 ; d2 = d3 >> 7 ; d2 = ..............a1 a0c1c0e1e0g1g0b1 b0d1d0f1f0h1h0a3 a2c3c2e3e2g3g2.. move.l d3,d2 ; 4 lsr.l #7,d2 ; 22 ; d1 = d3 & 55555555 ; d1 = ..a0..c0..e0..g0 ..b0..d0..f0..h0 ..a2..c2..e2..g2 ..b2..d2..f2..h2 move.l d3,d1 ; 4 and.l d6,d1 ; 8 d6=$55555555 ; d3 ^= d1 ; d3 = a1..c1..e1..g1.. b1..d1..f1..h1.. a3..c3..e3..g3.. b3..d3..f3..h3.. eor.l d1,d3 ; 8 ; d4 = d2 & 55555555 ; d4 = ..............a1 ..c1..e1..g1..b1 ..d1..f1..h1..a3 ..c3..e3..g3.... move.l d2,d4 ; 4 and.l d6,d4 ; 8 d6=$55555555 ; d2 ^= d4 ; d2 = ................ a0..c0..e0..g0.. b0..d0..f0..h0.. a2..c2..e2..g2.. eor.l d4,d2 ; 8 ; d3 = (d3 | d4) >> 1 ; d3 = ................ a1b1c1d1e1f1g1h1 ................ a3b3c3d3e3f3g3h3 or.l d4,d3 ; 8 lsr.l #1,d3 ; 10 ; d2 = d2 | d1 ; d2 = ................ a0b0c0d0e0f0g0h0 ................ a2b2c2d2e2f2g2h2 or.l d1,d2 ; 8 ; inner loop thus far is 220 move.b d3,(plane3)+ swap d3 move.b d3,(plane1)+ move.b d2,(plane2)+ swap d2 move.b d2,(plane0)+ subq.w #1,d0 ; check if end of line is reached bne.b mainloop ; no continue as usual ; --------------------------------------- addq.l #PlaneLineOffset,plane0 addq.l #PlaneLineOffset,plane1 addq.l #PlaneLineOffset,plane2 addq.l #PlaneLineOffset,plane3 subq.w #1,d5 ; tst.w a5 ; 68020+ instruction change to cmpa.w #0,a5 ; if you want 68000 compatibility bne.b PreMain ; ------------------------------------------------------------------------- FINC2P: movem.l (sp)+,d2-d7/a2-a4 rts ;----------------------------------------------------------------------------- end