Return to X86.S CVS log

Up to [UNIX Amiga Emulator] / uae

uae-0.6.0

/* These are optimized x86 assembly versions of pfield_linetoscr.
 * Feel free to send me Sparc/PPC/Alpha versions of this... :)
 */

#ifdef __DOS__
#define SYM(NAME) _##NAME
#define _color_regs _bpl_info
#else
#define SYM(NAME) NAME
#endif

/* Urgl... */
#define color_regs bpl_info

#ifndef __DOS__
	.local	lastcolor
	.comm	lastcolor,2
	.align 16
#else
	.comm	lastcolor,2
	.align 8
#endif
.globl pfield_linetoscr_full8
pfield_linetoscr_full8:
	pushl %ebp
	pushl %edi
	pushl %esi
	pushl %ebx
	movl 20(%esp),%esi
	movl 24(%esp),%edi
	
	xorl %edx,%edx
	movl SYM(diwfirstword),%eax
	cmpl $4,%eax
	jc .Loops1
	movl %edx,SYM(pixdata)-4(%eax)
	movl %edx,SYM(spixstate)-4(%eax)
.Loops1:	
	movl SYM(diwlastword),%eax
	movl %edx,SYM(pixdata)(%eax)
	movl %edx,SYM(spixstate)(%eax)
	
	movl SYM(acolors),%edx
	movb %dl,%dh
	movw %dx,%ax
	bswapl %eax
	movw %dx,%ax
	
	movl SYM(xlinebuffer),%ecx
	jmp .L71
.L8:
	cmpl 24(%esp),%esi
	jge .L5
	movl %eax,(%ecx)
	addl $4,%ecx
	movl %edx,%esi
.L71:
	leal 4(%esi),%edx
	cmpl %edx,SYM(diwfirstword)
	jg .L8
.L5:
        movl %ecx,SYM(xlinebuffer)

        cmpl $0,SYM(bplham)
	je .L9
	cmpl $6,SYM(bplplanecnt)
	jne .L9
	
	/* HAM */
	/* This is invalid in pfield_linetoscr, but it's OK here */
	movw SYM(color_regs),%ax
	movw %ax,lastcolor
.L73:
	cmpl %esi,SYM(diwlastword)
	jle .L27
	cmpl 24(%esp),%esi
	jge .L27
	movzbl SYM(pixdata)(%esi),%eax
	movl %eax,%edx
	andl $48,%edx
	cmpl $16,%edx
	je .L16
	jg .L21
	testl %edx,%edx
	je .L15
	jmp .L14
.L21:
	cmpl $32,%edx
	je .L17
	cmpl $48,%edx
	je .L18
	jmp .L14
.L15:
	movw SYM(color_regs)(,%eax,2),%ax
	movw %ax,lastcolor
	jmp .L14
.L16:
	movw lastcolor,%dx
	andw $4080,%dx
	andw $15,%ax
	jmp .L72
.L17:
	movzbw lastcolor,%dx
	andw $15,%ax
	salw $8,%ax
	jmp .L72
.L18:
	movw lastcolor,%dx
	andw $3855,%dx
	andw $15,%ax
	salw $4,%ax
.L72:
	orw %ax,%dx
	movw %dx,lastcolor
.L14:
	cmpb $0,SYM(spixstate)(%esi)
	je .L22
	movsbl SYM(spixels)(%esi),%eax
	movl SYM(acolors)+64(,%eax,4),%edx
	movl SYM(xlinebuffer),%eax
	movb %dl,(%eax)
	incl SYM(xlinebuffer)
	movb $0,SYM(spixstate)(%esi)
	movb $0,SYM(spixels)(%esi)
	jmp .L24
.L22:
	movzwl lastcolor,%eax
	movl SYM(xcolors)(,%eax,4),%edx
	movl SYM(xlinebuffer),%eax
	movb %dl,(%eax)
	incl SYM(xlinebuffer)
.L24:
	incl %esi
	jmp .L73
.L9:
	cmpl $0,SYM(bpldualpf)
	je .L74
	
	/* Dual Playfield */
	
	movl $SYM(dblpf_ind1),%ebp
	movl $SYM(dblpf_2nd1),%edi
	cmpl $0,SYM(bpldualpfpri)
	je .L29
	movl $SYM(dblpf_ind2),%ebp
	movl $SYM(dblpf_2nd2),%edi
.L29:
	cmpl %esi,SYM(diwlastword)
	jle .L27
	
	cmpl 24(%esp),%esi
	jge .L27

	xorl %edx,%edx
	movl SYM(pixdata)(%esi),%eax
	movl SYM(spixstate)(%esi),%ebx
	testl %ebx,%ebx
	jne .L101
	
	/* no sprites */

	movb %al,%dl
	movb (%ebp,%edx,4),%dl
	movb SYM(acolors)(,%edx,4),%cl
	movb %ah,%dl
	movb (%ebp,%edx,4),%dl
	movb SYM(acolors)(,%edx,4),%ch
	bswapl %eax
	bswapl %ecx
	movb %al,%dl
	movb (%ebp,%edx,4),%dl
	movb SYM(acolors)(,%edx,4),%cl
	movb %ah,%dl
	movb (%ebp,%edx,4),%dl
	movb SYM(acolors)(,%edx,4),%ch
	bswapl %ecx
	
	movl SYM(xlinebuffer),%eax
	movl %ecx,(%eax)
	addl $4,SYM(xlinebuffer)
	addl $4,%esi
	jmp .L29
	
.L101:

        /* dual playfield with sprites */

	testb %bl,%bl
	je .La3
	movb %al,%dl
	movl (%edi,%edx,4),%edx
	testl %edx,%edx
	je .La2
	pushl %ebp
	movl %edx,%ebp
	movb %bl,%dl
	cmpl plfpri(,%ebp,4),%edx
	popl %ebp
	jge .La3
.La2:
	movb SYM(spixels)(%esi),%dl
	movb SYM(acolors)+64(,%edx,4),%cl
	jmp .La1
.La3:
	movb %al,%dl
	movb (%ebp,%edx,4),%dl
	movb SYM(acolors)(,%edx,4),%cl
.La1:

	testb %bh,%bh
	je .La6
	movb %ah,%dl
	movl (%edi,%edx,4),%edx
	testl %edx,%edx
	je .La5
	pushl %ebp
	movl %edx,%ebp
	movb %bh,%dl
	cmpl plfpri(,%ebp,4),%edx
	popl %ebp
	jge .La6
.La5:
	movb SYM(spixels)+1(%esi),%dl
	movb SYM(acolors)+64(,%edx,4),%ch
	jmp .La4
.La6:
	movb %ah,%dl
	movb (%ebp,%edx,4),%dl
	movb SYM(acolors)(,%edx,4),%ch
.La4:
	bswapl %ecx	
	bswapl %eax
	bswapl %ebx
	
	testb %bl,%bl
	je .La9
	movb %al,%dl
	movl (%edi,%edx,4),%edx
	testl %edx,%edx
	je .La8
	pushl %ebp
	movl %edx,%ebp
	movb %bl,%dl
	cmpl plfpri(,%ebp,4),%edx
	popl %ebp
	jge .La9
.La8:
	movb SYM(spixels)+3(%esi),%dl
	movb SYM(acolors)+64(,%edx,4),%cl
	jmp .La7
.La9:
	movb %al,%dl
	movb (%ebp,%edx,4),%dl
	movb SYM(acolors)(,%edx,4),%cl
.La7:

	testb %bh,%bh
	je .La12
	movb %ah,%dl
	movl (%edi,%edx,4),%edx
	testl %edx,%edx
	je .La11
	pushl %ebp
	movl %edx,%ebp
	movb %bh,%dl
	cmpl plfpri(,%ebp,4),%edx
	popl %ebp
	jge .La12
.La11:
	movb SYM(spixels)+2(%esi),%dl
	movb SYM(acolors)+64(,%edx,4),%ch
	jmp .La10
.La12:
	movb %ah,%dl
	movb (%ebp,%edx,4),%dl
	movb SYM(acolors)(,%edx,4),%ch
.La10:
        movl $0,SYM(spixstate)(%esi)
        movl $0,SYM(spixels)(%esi)
	addl $4,%esi
	
	bswapl %ecx

	movl SYM(xlinebuffer),%eax
	movl %ecx,(%eax)
	addl $4,SYM(xlinebuffer)
	
	jmp .L29
	
	/* normal case */
	
.L65:
	cmpl 24(%esp),%esi
	jge .L27
	
	xorl %edx,%edx
	movl SYM(pixdata)(%esi),%eax
	movl SYM(spixstate)(%esi),%ebx
	testl %ebx,%ebx
	jne .L100
	
	/* no sprites */
	
	movb %al,%dl
	movb SYM(acolors)(,%edx,4),%cl
	movb %ah,%dl
	movb SYM(acolors)(,%edx,4),%ch
	bswapl %eax
	bswapl %ecx
	movb %al,%dl
	movb SYM(acolors)(,%edx,4),%cl
	movb %ah,%dl
	movb SYM(acolors)(,%edx,4),%ch
	bswapl %ecx
	
	movl SYM(xlinebuffer),%eax
	movl %ecx,(%eax)
	addl $4,SYM(xlinebuffer)
	addl $4,%esi
	jmp .L74
	
.L100:
        /* Here, we have to deal with sprites, too. */
	
	testb %bl,%bl
	je .L46
	testb %al,%al
	je .L47
	movb %bl,%dl
	cmpl SYM(plfpri)+8,%edx
	jge .L46
.L47:
	movb SYM(spixels)(%esi),%dl
	movb SYM(acolors)+64(,%edx,4),%cl
	jmp .L75
.L46:
	movb %al,%dl
	movb SYM(acolors)(,%edx,4),%cl
.L75:

	testb %bh,%bh
	je .L51
	testb %ah,%ah
	je .L52
	movb %bh,%dl
	cmpl SYM(plfpri)+8,%edx
	jge .L51
	
.L52:
	movb SYM(spixels)+1(%esi),%dl
	movb SYM(acolors)+64(,%edx,4),%ch
	jmp .L76
.L51:
	movb %ah,%dl
	movb SYM(acolors)(,%edx,4),%ch
.L76:	
	
	bswapl %ebx
	bswapl %ecx
	bswapl %eax
	
	testb %bl,%bl
	je .L56
	testb %al,%al
	je .L55
	movb %bl,%dl	
	cmpl SYM(plfpri)+8,%edx
	jge .L56
.L55:
	movb SYM(spixels)+3(%esi),%dl
	movb SYM(acolors)+64(,%edx,4),%cl
	jmp .L57
.L56:
	movb %al,%dl
	movb SYM(acolors)(,%edx,4),%cl
.L57:

	testb %bh,%bh
	je .L58
	testb %ah,%ah
	je .L59
	movb %bh,%dl
	cmpl SYM(plfpri)+8,%edx
	jge .L58
.L59:
	movb SYM(spixels)+2(%esi),%dl
	movb SYM(acolors)+64(,%edx,4),%ch
	jmp .L60
.L58:
	movb %ah,%dl
	movb SYM(acolors)(,%edx,4),%ch
.L60:
        movl $0,SYM(spixstate)(%esi)
        movl $0,SYM(spixels)(%esi)
	addl $4,%esi
	
	bswapl %ecx

	movl SYM(xlinebuffer),%eax
	movl %ecx,(%eax)
	addl $4,SYM(xlinebuffer)
.L74:
	cmpl %esi,SYM(diwlastword)
	jg .L65
.L27:

	/* Finish off the line */
	
	movl SYM(acolors),%edx
	movb %dl,%dh
	movw %dx,%ax
	bswapl %eax
	movw %dx,%ax
	
	movl SYM(xlinebuffer),%ecx
	movl 24(%esp),%edi
.L70:
	cmpl %edi,%esi
	jge .L66
	
	movl %eax,(%ecx)
	addl $4,%ecx
	addl $4,%esi
	jmp  .L70
.L66:
        movl %ecx,SYM(xlinebuffer)
	
	popl %ebx
	popl %esi
	popl %edi
	popl %ebp
	ret
.Lfe2:

#ifndef __DOS__
	.align 16
#else
	.align 8
#endif
.globl pfield_linetoscr_full16
pfield_linetoscr_full16:
        pushl %edi
	pushl %esi
	pushl %ebx
	pushl %ebp
	
	xorl %edx,%edx
	movl SYM(diwfirstword),%eax
	cmpl $4,%eax
	jc .Loops2
	movl %edx,SYM(pixdata)-4(%eax)
	movl %edx,SYM(spixstate)-4(%eax)
.Loops2:
	movl SYM(diwlastword),%eax
	movl %edx,SYM(pixdata)(%eax)
	movl %edx,SYM(spixstate)(%eax)
	
	movl 20(%esp),%esi
	
	movw SYM(acolors),%ax
	sall $16,%eax
	movw SYM(acolors),%ax
	
	movl SYM(xlinebuffer),%ecx
	jmp .L2443
.L2447:
	cmpl 24(%esp),%esi
	jge .L2400
	movl %eax,(%ecx)
	movl %eax,4(%ecx)
	addl $8,%ecx
	movl %edx,%esi
.L2443:
	leal 4(%esi),%edx
	cmpl %edx,SYM(diwfirstword)
	jg .L2447
.L2400:
        movl %ecx,SYM(xlinebuffer)
	
	cmpl $0,SYM(bplham)
	je .L2448
	cmpl $6,SYM(bplplanecnt)
	jne .L2448
	
	/* HAM */
	
	/* This is invalid in pfield_linetoscr, but it's OK here */
	movw SYM(color_regs),%ax
	movw %ax,lastcolor
.L2499:
	cmpl %esi,SYM(diwlastword)
	jle .L2027
	cmpl 24(%esp),%esi
	jge .L2494
	movzbl SYM(pixdata)(%esi),%eax
	movl %eax,%edx
	andl $48,%edx
	cmpl $16,%edx
	je .L2455
	jg .L2460
	testl %edx,%edx
	je .L2454
	jmp .L2453
.L2460:
	cmpl $32,%edx
	je .L2456
	cmpl $48,%edx
	je .L2457
	jmp .L2453
.L2454:
	movw SYM(color_regs)(,%eax,2),%ax
	movw %ax,lastcolor
	jmp .L2453
.L2455:
	movw lastcolor,%dx
	andw $4080,%dx
	andw $15,%ax
	jmp .L2498
.L2456:
	movzbw lastcolor,%dx
	andw $15,%ax
	salw $8,%ax
	jmp .L2498
.L2457:
	movw lastcolor,%dx
	andw $3855,%dx
	andw $15,%ax
	salw $4,%ax
.L2498:
	orw %ax,%dx
	movw %dx,lastcolor
.L2453:
	cmpb $0,SYM(spixstate)(%esi)
	je .L2461
	movsbl SYM(spixels)(%esi),%eax
	movl SYM(acolors)+64(,%eax,4),%edx
	movl SYM(xlinebuffer),%eax
	movw %dx,(%eax)
	addl $2,SYM(xlinebuffer)
	movb $0,SYM(spixstate)(%esi)
	movb $0,SYM(spixels)(%esi)
	jmp .L2463
.L2461:
	movzwl lastcolor,%eax
	movl SYM(xcolors)(,%eax,4),%edx
	movl SYM(xlinebuffer),%eax
	movw %dx,(%eax)
	addl $2,SYM(xlinebuffer)
.L2463:
	incl %esi
	jmp .L2499
.L2448:
        cmpl $0,SYM(bpldualpf)
	je .L2500
	
	/* Dual Playfield */
	
	xorl %edx,%edx
	
	movl $SYM(dblpf_ind1),%ebp
	movl $SYM(dblpf_2nd1),%edi
	cmpl $0,SYM(bpldualpfpri)
	je .L2468
	movl $SYM(dblpf_ind2),%ebp
	movl $SYM(dblpf_2nd2),%edi
.L2468:
	cmpl %esi,SYM(diwlastword)
	jle .L2027
	cmpl 24(%esp),%esi
	jge .L2494
	
	movl SYM(pixdata)(%esi),%eax
	movl SYM(spixstate)(%esi),%ebx
	testl %ebx,%ebx
	jne .Lb0
	
	/* Dual Playfield, no sprites */
	
	movl SYM(xlinebuffer),%ebx
	
	movb %ah,%dl
	movb (%ebp,%edx,4),%dl
	movw SYM(acolors)(,%edx,4),%cx
	sall $16,%ecx
	movb %al,%dl
	movb (%ebp,%edx,4),%dl
	movw SYM(acolors)(,%edx,4),%cx
	movl %ecx,(%ebx)
	
	bswapl %eax
	movb %al,%dl
	movb (%ebp,%edx,4),%dl
	movw SYM(acolors)(,%edx,4),%cx
	sall $16,%ecx
	movb %ah,%dl
	movb (%ebp,%edx,4),%dl
	movw SYM(acolors)(,%edx,4),%cx
	movl %ecx,4(%ebx)
	addl $8,SYM(xlinebuffer)
	addl $4,%esi
	jmp .L2468
	
	/* Dual Playfield with sprites */
	
.Lb0:		
	testb %bh,%bh
	je .Lb3
	movb %ah,%dl
	movl (%edi,%edx,4),%edx
	testl %edx,%edx
	je .Lb2
	pushl %ebp
	movl %edx,%ebp
	movb %bh,%dl
	cmpl plfpri(,%ebp,4),%edx
	popl %ebp
	jge .Lb3
.Lb2:
	movb SYM(spixels)+1(%esi),%dl
	movw SYM(acolors)+64(,%edx,4),%cx
	jmp .Lb1
.Lb3:
	movb %ah,%dl
	movb (%ebp,%edx,4),%dl
	movw SYM(acolors)(,%edx,4),%cx
.Lb1:
        sall $16,%ecx

	testb %bl,%bl
	je .Lb6
	movb %al,%dl
	movl (%edi,%edx,4),%edx
	testl %edx,%edx
	je .Lb5
	pushl %ebp
	movl %edx,%ebp
	movb %bl,%dl
	cmpl plfpri(,%ebp,4),%edx
	popl %ebp
	jge .Lb6
.Lb5:
	movb SYM(spixels)(%esi),%dl
	movw SYM(acolors)+64(,%edx,4),%cx
	jmp .Lb4
.Lb6:
	movb %al,%dl
	movb (%ebp,%edx,4),%dl
	movw SYM(acolors)(,%edx,4),%cx
.Lb4:
	pushl %eax
	movl SYM(xlinebuffer),%eax
	movl %ecx,(%eax)
	popl %eax
	
	bswapl %ebx
	bswapl %eax
	
	testb %bl,%bl
	je .Lb9
	movb %al,%dl
	movl (%edi,%edx,4),%edx
	testl %edx,%edx
	je .Lb8
	pushl %ebp
	movl %edx,%ebp
	movb %bl,%dl
	cmpl plfpri(,%ebp,4),%edx
	popl %ebp
	jge .Lb9
.Lb8:
	movb SYM(spixels)+3(%esi),%dl
	movw SYM(acolors)+64(,%edx,4),%cx
	jmp .Lb7
.Lb9:
	movb %al,%dl
	movb (%ebp,%edx,4),%dl
	movw SYM(acolors)(,%edx,4),%cx
.Lb7:
        sall $16,%ecx

	testb %bh,%bh
	je .Lb12
	movb %ah,%dl
	movl (%edi,%edx,4),%edx
	testl %edx,%edx
	je .Lb11
	pushl %ebp
	movl %edx,%ebp
	movb %bh,%dl
	cmpl plfpri(,%ebp,4),%edx
	popl %ebp
	jge .Lb12
.Lb11:
	movb SYM(spixels)+2(%esi),%dl
	movw SYM(acolors)+64(,%edx,4),%cx
	jmp .Lb10
.Lb12:
	movb %ah,%dl
	movb (%ebp,%edx,4),%dl
	movw SYM(acolors)(,%edx,4),%cx
.Lb10:
        movl $0,SYM(spixstate)(%esi)
        movl $0,SYM(spixels)(%esi)
	addl $4,%esi
	
	movl SYM(xlinebuffer),%eax
	movl %ecx,4(%eax)
	addl $8,SYM(xlinebuffer)
	
	jmp .L2468
#if 0
	
	movsbl SYM(spixels)(%esi),%eax
	movl SYM(acolors)+64(,%eax,4),%edx
	movl SYM(xlinebuffer),%eax
	movw %dx,(%eax)
	addl $2,SYM(xlinebuffer)
	movb $0,SYM(spixstate)(%esi)
	movb $0,SYM(spixels)(%esi)
	jmp .L2476
.L2474:
	movzbl SYM(pixdata)(%esi),%eax
	movl (%ebp,%eax,4),%eax
	movl SYM(acolors)(,%eax,4),%edx
	movl SYM(xlinebuffer),%eax
	movw %dx,(%eax)
	addl $2,SYM(xlinebuffer)
.L2476:
	incl %esi
	jmp .L2468
#endif

	/* Normal case */
.L2492:
	cmpl 24(%esp),%esi
	jge .L2027
	
	xorl %edx,%edx
	movl SYM(pixdata)(%esi),%eax
	movl SYM(spixstate)(%esi),%ebx
	testl %ebx,%ebx
	jne .L20100
	
	/* no sprites */
	
	movl SYM(xlinebuffer),%ebp
	
	movb %ah,%dl
	movw SYM(acolors)(,%edx,4),%cx
	sall $16,%ecx
	movb %al,%dl
	movw SYM(acolors)(,%edx,4),%cx
	movl %ecx,(%ebp)

        bswapl %eax
	
	movb %al,%dl
	movw SYM(acolors)(,%edx,4),%cx
	sall $16,%ecx
	movb %ah,%dl
	movw SYM(acolors)(,%edx,4),%cx
	movl %ecx,4(%ebp)
	
	addl $8,%ebp
	movl %ebp,SYM(xlinebuffer)
	addl $4,%esi
	jmp .L2500
	
.L20100:
        /* Here, we have to deal with sprites, too. */
	
	movl SYM(xlinebuffer),%ebp

	testb %bh,%bh
	je .L2051
	testb %ah,%ah
	je .L2052
	movb %bh,%dl
	cmpl SYM(plfpri)+8,%edx
	jge .L2051
	
.L2052:
	movb SYM(spixels)+1(%esi),%dl
	movw SYM(acolors)+64(,%edx,4),%cx
	jmp .L2076
.L2051:
	movb %ah,%dl
	movw SYM(acolors)(,%edx,4),%cx
.L2076:	
        sall $16,%ecx
	
	testb %bl,%bl
	je .L2046
	testb %al,%al
	je .L2047
	movb %bl,%dl
	cmpl SYM(plfpri)+8,%edx
	jge .L2046
.L2047:
	movb SYM(spixels)(%esi),%dl
	movw SYM(acolors)+64(,%edx,4),%cx
	jmp .L2075
.L2046:
	movb %al,%dl
	movw SYM(acolors)(,%edx,4),%cx
.L2075:
	
	bswapl %ebx
	bswapl %eax
	movl %ecx,(%ebp)
	
	testb %bl,%bl
	je .L2056
	testb %al,%al
	je .L2055
	movb %bl,%dl
	cmpl SYM(plfpri)+8,%edx
	jge .L2056
.L2055:
	movb SYM(spixels)+3(%esi),%dl
	movw SYM(acolors)+64(,%edx,4),%cx
	jmp .L2057
.L2056:
	movb %al,%dl
	movw SYM(acolors)(,%edx,4),%cx
.L2057:
        sall $16,%ecx

        testb %bh,%bh
	je .L2058
	testb %ah,%ah
	je .L2059
	movb %bh,%dl
	cmpl SYM(plfpri)+8,%edx
	jge .L2058
.L2059:
	movb SYM(spixels)+2(%esi),%dl
	movw SYM(acolors)+64(,%edx,4),%cx
	jmp .L2060
.L2058:
	movb %ah,%dl
	movw SYM(acolors)(,%edx,4),%cx
.L2060:

        movl %ecx,4(%ebp)
        movl $0,SYM(spixstate)(%esi)
        movl $0,SYM(spixels)(%esi)
	addl $4,%esi

	addl $8,SYM(xlinebuffer)
.L2500:
	cmpl %esi,SYM(diwlastword)
	jg .L2492
	
	/* Finish off the line */
	
.L2027:
        movl 24(%esp),%edi
	cmpl %edi,%esi
	jnl .L2494
	movl SYM(acolors),%edx
	movl SYM(xlinebuffer),%eax
.L2495:
	incl %esi
	movw %dx,(%eax)
	addl $2,%eax

	cmpl %edi,%esi
	jl .L2495
	movl %eax,SYM(xlinebuffer)
.L2494:
        popl %ebp
	popl %ebx
	popl %esi
	popl %edi
	ret
.Lfe21:

        .ident	"Crux CC 3.14.15"

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.