|
|
MiNT 0.95 pl13
; ; quickmove(char *dst, char *src, long nbytes): ; quickly copy "nbytes" bytes from src to dst. Assumes that both ; src and dst are word aligned. ; SECTION TEXT XDEF _quickmove _quickmove: move.l 4(sp),a0 ; get dst move.l 8(sp),a1 ; get src move.l 12(sp),d0 ; get nbytes lsr.l #8,d0 ; lsr.l #1,d0 ; d0 = nbytes / 512 subq.l #1,d0 ; prepare for dbra loop bmi Leftover ; if < 0, skip movem.l d1-d7/a2-a6,-(sp) ; save regs L1: movem.l (a1)+,d1-d7/a2-a6 ; read 12*4 = 48 bytes movem.l d1-d7/a2-a6,(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 2nd read movem.l d1-d7/a2-a6,48(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 3rd read movem.l d1-d7/a2-a6,96(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 4th read movem.l d1-d7/a2-a6,144(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 5th movem.l d1-d7/a2-a6,192(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 6th movem.l d1-d7/a2-a6,240(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 7th movem.l d1-d7/a2-a6,288(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 8th movem.l d1-d7/a2-a6,336(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 9th movem.l d1-d7/a2-a6,384(a0) ; movem.l (a1)+,d1-d7/a2-a6 ; 10th movem.l d1-d7/a2-a6,432(a0) ; At this point, 480 bytes done movem.l (a1)+,d1-d7/a2 ; Only do 32 more bytes movem.l d1-d7/a2,480(a0) ; for a total of 512 lea 512(a0),a0 subq.l #1,d0 bge.s L1 movem.l (sp)+,d1-d7/a2-a6 ; pop registers Leftover: ; do the remaining bytes move.l 12(sp),d1 and.l #$01ff,d1 ; d1 = nbytes % 512 subq.l #1,d1 ; prepare for dbra loop bmi.s Ldone L2: move.b (a1)+,(a0)+ dbra d1,L2 Ldone: rts ; return END
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.