DOS updates from Daniel Borca.

1 files changed, 363 insertions, 106 deletions

diff --git a/src/mesa/drivers/dos/virtual.S b/src/mesa/drivers/dos/virtual.S
index 3633161c982..4dbbbb3e3f8 100644
--- a/src/mesa/drivers/dos/virtual.S
+++ b/src/mesa/drivers/dos/virtual.S
@@ -23,7 +23,7 @@
  */

 

 /*

- * DOS/DJGPP device driver v1.1 for Mesa 4.0

+ * DOS/DJGPP device driver v1.3 for Mesa 5.0

  *

  *  Copyright (C) 2002 - Borca Daniel

  *  Email : [email protected]

@@ -34,60 +34,153 @@
 		.file	"virtual.S"

 

 /*

- * extern word32 vl_hw_granularity;

- * extern int vl_video_selector;

- * extern int vl_current_offset, vl_current_delta;

+ * extern void *vl_current_draw_buffer;

+ * extern int vl_current_width, vl_current_bytes;

  */

 

 		.text

 

-/*

- * void v_clear16 (void *buffer, int bytes, int color);

+/* Desc: void v_clear8 (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

  */

-		.balign	4

+		.p2align 5,,31

+		.global	_v_clear8

+_v_clear8:

+		movl	4(%esp), %eax

+		movb	%al, %ah

+		pushw	%ax

+		pushw	%ax

+		popl	%eax

+		jmp	_v_clear_common

+

+/* Desc: void v_clear16 (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

+ */

+		.p2align 5,,31

 		.global	_v_clear16

 _v_clear16:

-		movl	12(%esp), %eax

+		movl	4(%esp), %eax

 		pushw	%ax

 		pushw	%ax

 		popl	%eax

 		jmp	_v_clear_common

 

-/*

- * void v_clear32 (void *buffer, int bytes, int color);

+/* Desc: void v_clear32 (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

  */

-		.balign	4

+		.p2align 5,,31

 		.global	_v_clear32

 _v_clear32:

-		movl	12(%esp), %eax

+		movl	4(%esp), %eax

 		.balign	4

 _v_clear_common:

-		movl	8(%esp), %ecx

-		movl	4(%esp), %edx

+		movl	_vl_current_bytes, %ecx

+		movl	_vl_current_draw_buffer, %edx

 		shrl	$2, %ecx

-	0:

 		.balign	4

+	0:

 		movl	%eax, (%edx)

 		addl	$4, %edx

 		decl	%ecx

 		jnz	0b

 		ret

 

-/*

- * void v_clear24 (void *buffer, int bytes, int color);

+/* Desc: void v_clear8_mmx (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

+ */

+		.p2align 5,,31

+		.global	_v_clear8_mmx

+_v_clear8_mmx:

+		movl	4(%esp), %eax

+		movb	%al, %ah

+		pushw	%ax

+		pushw	%ax

+		pushw	%ax

+		pushw	%ax

+		jmp	_v_clear_common_mmx

+

+/* Desc: void v_clear16_mmx (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

  */

+		.p2align 5,,31

+		.global	_v_clear16_mmx

+_v_clear16_mmx:

+		movl	4(%esp), %eax

+		pushw	%ax

+		pushw	%ax

+		pushw	%ax

+		pushw	%ax

+		jmp	_v_clear_common_mmx

+

+/* Desc: void v_clear32_mmx (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

+ */

+		.p2align 5,,31

+		.global	_v_clear32_mmx

+_v_clear32_mmx:

+		movl	4(%esp), %eax

+		pushl	%eax

+		pushl	%eax

 		.balign	4

+_v_clear_common_mmx:

+#ifdef USE_MMX_ASM

+		movl	_vl_current_bytes, %ecx

+		movl	_vl_current_draw_buffer, %edx

+		movq	(%esp), %mm0

+		shrl	$3, %ecx

+		.balign	4

+	0:

+		movq	%mm0, (%edx)

+		addl	$8, %edx

+		decl	%ecx

+		jnz	0b

+		emms

+#endif

+		addl	$8, %esp

+		ret

+

+/* Desc: void v_clear24 (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

+ */

+		.p2align 5,,31

 		.global	_v_clear24

 _v_clear24:

-		movl	8(%esp), %edx

 		movl	$0xaaaaaaab, %eax

-		mull	%edx

-		movl	12(%esp), %eax

+		mull	_vl_current_bytes

+		movl	4(%esp), %eax

 		movl	%edx, %ecx

-		movl	4(%esp), %edx

 		pushl	%ebx

+		movl	_vl_current_draw_buffer, %edx

 		shrl	%ecx

-		movb	18(%esp), %bl

+		movb	10(%esp), %bl

 		.balign	4

 	0:

 		movw	%ax, (%edx)

@@ -98,140 +191,304 @@ _v_clear24:
 		popl	%ebx

 		ret

 

-/*

- * void b_dump_virtual (void *buffer, int stride, int height);

+/* Desc: void v_clear24_mmx (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

  */

+		.p2align 5,,31

+		.global	_v_clear24_mmx

+_v_clear24_mmx:

+#ifdef USE_MMX_ASM

+		movl	4(%esp), %eax

+		movl	%eax, %edx

+		movl	%eax, %ecx

+		shll	$16, %edx

+		rorl	$8, %ecx

+		movw	%cx, %dx

+		rorl	$16, %ecx

+		movb	%dh, %cl

+		shll	$8, %eax

+		movb	%ch, %al

+		rorl	$8, %eax

+

+		pushl	%edx

+		pushl	%eax

+		movq	(%esp), %mm0

+		pushl	%ecx

+		movq	(%esp), %mm1

+		pushl	%edx

+		movq	(%esp), %mm2

+

+		movl	$0xaaaaaaab, %eax

+		mull	_vl_current_bytes

+		movl	%edx, %ecx

+		movl	_vl_current_draw_buffer, %edx

+		shrl	$4, %ecx

 		.balign	4

-		.global	_b_dump_virtual

-_b_dump_virtual:

-		pushl	%ebx

+	0:

+		movq	%mm0, (%edx)

+		movq	%mm1, 8(%edx)

+		movq	%mm2, 16(%edx)

+		addl	$24, %edx

+		decl	%ecx

+		jnz	0b

+		emms

+		addl	$16, %esp

+#endif

+		ret

+

+/* Desc: void v_rect8 (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

+ */

+		.p2align 5,,31

+		.global	_v_rect8

+_v_rect8:

+		cld

 		pushl	%esi

 		pushl	%edi

-		pushl	%ebp

-		movl	_vl_video_selector, %fs

-		movl	4*4+4+0(%esp), %esi

-		movl	_vl_hw_granularity, %ebp

-		xorl	%edx, %edx

-		movl	_vl_current_offset, %eax

-		divl	%ebp

-		movl	%edx, %edi

-		pushl	%eax

-		movl	%eax, %edx

-		xorl	%ebx, %ebx

-		movw	$0x4f05, %ax

-		int	$0x10

-		movl	_vl_current_delta, %ebx

-		movl	5*4+4+4(%esp), %ecx

-		movl	5*4+4+8(%esp), %edx

-		shrl	$2, %ecx

+		movl	28(%esp), %eax

+		movl	_vl_current_width, %esi

+		movl	16(%esp), %edi

+		movb	%al, %ah

+		movl	20(%esp), %ecx

+		imull	%esi, %edi

+		movl	24(%esp), %edx

+		subl	%ecx, %esi

+		addl	12(%esp), %edi

+		pushw	%ax

+		pushw	%ax

+		pushl	%ds

+		popl	%es

+		addl	_vl_current_draw_buffer, %edi

+		popl	%eax

+		pushl	%ebx

+		movl	%ecx, %ebx

+		andl	$3, %ebx

 		.balign	4

 	0:

 		pushl	%ecx

 		.balign	4

 	1:

-		cmpl	%ebp, %edi

-		jb	2f

-		pushl	%ebx

-		pushl	%edx

-		incl	12(%esp)

-		movw	$0x4f05, %ax

-		movl	12(%esp), %edx

-		xorl	%ebx, %ebx

-		int	$0x10

-		popl	%edx

+		shrl	$2, %ecx

+		rep;	stosl

+		testl	%ebx, %ebx

+		jz	2f

+		movl	%ebx, %ecx

+		rep;	stosb

+		.balign	4

+	2:

+		popl	%ecx

+		addl	%esi, %edi

+		decl	%edx

+		jnz	0b

 		popl	%ebx

-		subl	%ebp, %edi

+		popl	%edi

+		popl	%esi

+		ret

+

+/* Desc: void v_rect16 (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

+ */

+		.p2align 5,,31

+		.global	_v_rect16

+_v_rect16:

+		cld

+		pushl	%esi

+		pushl	%edi

+		movl	28(%esp), %eax

+		movl	_vl_current_width, %esi

+		movl	16(%esp), %edi

+		movl	20(%esp), %ecx

+		imull	%esi, %edi

+		movl	24(%esp), %edx

+		subl	%ecx, %esi

+		addl	12(%esp), %edi

+		pushw	%ax

+		shll	%esi

+		pushw	%ax

+		shll	%edi

+		pushl	%ds

+		popl	%es

+		addl	_vl_current_draw_buffer, %edi

+		popl	%eax

+		.balign	4

+	0:

+		pushl	%ecx

+		.balign	4

+	1:

+		shrl	%ecx

+		rep;	stosl

+		jnc	2f

+		stosw

+		.balign	4

 	2:

-		movl	(%esi), %eax

-		addl	$4, %esi

-		movl	%eax, %fs:(%edi)

-		addl	$4, %edi

+		popl	%ecx

+		addl	%esi, %edi

+		decl	%edx

+		jnz	0b

+		popl	%edi

+		popl	%esi

+		ret

+

+/* Desc: void v_rect24 (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

+ */

+		.p2align 5,,31

+		.global	_v_rect24

+_v_rect24:

+		pushl	%esi

+		pushl	%edi

+		movl	28(%esp), %eax

+		movl	_vl_current_width, %esi

+		movl	16(%esp), %edi

+		movl	20(%esp), %ecx

+		imull	%esi, %edi

+		movl	24(%esp), %edx

+		subl	%ecx, %esi

+		addl	12(%esp), %edi

+		leal	(%esi, %esi, 2), %esi

+		pushl	%ebx

+		leal	(%edi, %edi, 2), %edi

+		movl	%eax, %ebx

+		addl	_vl_current_draw_buffer, %edi

+		shrl	$16, %ebx

+		.balign	4

+	0:

+		pushl	%ecx

+		.balign	4

+	1:

+		movw	%ax, (%edi)

+		movb	%bl, 2(%edi)

+		addl	$3, %edi

 		decl	%ecx

 		jnz	1b

 		popl	%ecx

-		addl	%ebx, %edi

+		addl	%esi, %edi

 		decl	%edx

 		jnz	0b

-		popl	%eax

-		popl	%ebp

+		popl	%ebx

 		popl	%edi

 		popl	%esi

-		popl	%ebx

 		ret

 

-/*

- * void l_dump_virtual (void *buffer, int stride, int height);

+/* Desc: void v_rect32 (int color);

+ *

+ * In  : color

+ * Out : -

+ *

+ * Note: uses current draw buffer

  */

-		.balign	4

-		.global	_l_dump_virtual

-_l_dump_virtual:

-		pushl	%ebx

+		.p2align 5,,31

+		.global	_v_rect32

+_v_rect32:

 		pushl	%esi

 		pushl	%edi

-		movl	_vl_video_selector, %fs

-		movl	3*4+4+0(%esp), %esi

-		movl	_vl_current_offset, %edi

-		movl	3*4+4+4(%esp), %ecx

-		movl	3*4+4+8(%esp), %edx

-		movl	_vl_current_delta, %ebx

-		shrl	$2, %ecx

+		movl	_vl_current_width, %esi

+		movl	16(%esp), %edi

+		movl	20(%esp), %ecx

+		imull	%esi, %edi

+		movl	24(%esp), %edx

+		subl	%ecx, %esi

+		addl	12(%esp), %edi

+		shll	$2, %esi

+		shll	$2, %edi

+		movl	28(%esp), %eax

+		addl	_vl_current_draw_buffer, %edi

 		.balign	4

 	0:

 		pushl	%ecx

 		.balign	4

 	1:

-		movl	(%esi), %eax

-		addl	$4, %esi

-		movl	%eax, %fs:(%edi)

+		movl	%eax, (%edi)

 		addl	$4, %edi

 		decl	%ecx

 		jnz	1b

 		popl	%ecx

-		addl	%ebx, %edi

+		addl	%esi, %edi

 		decl	%edx

 		jnz	0b

 		popl	%edi

 		popl	%esi

-		popl	%ebx

 		ret

 

-/*

- * void v_putpixel16 (void *buffer, int offset, int color);

+/* Desc: void v_putpixel8 (unsigned int offset, int color);

+ *

+ * In  : offset within buffer, color

+ * Out : -

+ *

+ * Note: uses current draw buffer

  */

-		.balign	4

+		.p2align 5,,31

+		.global	_v_putpixel8

+_v_putpixel8:

+		movl	8(%esp), %ecx

+		movl	4(%esp), %edx

+		movl	_vl_current_draw_buffer, %eax

+		movb	%cl, (%eax,%edx)

+		ret

+

+/* Desc: void v_putpixel16 (unsigned int offset, int color);

+ *

+ * In  : offset within buffer, color

+ * Out : -

+ *

+ * Note: uses current draw buffer

+ */

+		.p2align 5,,31

 		.global	_v_putpixel16

 _v_putpixel16:

-		movl	8(%esp), %edx

-		shll	%edx

-		movl	12(%esp), %eax

-		addl	4(%esp), %edx

-		movw	%ax, (%edx)

+		movl	8(%esp), %ecx

+		movl	4(%esp), %edx

+		movl	_vl_current_draw_buffer, %eax

+		movw	%cx, (%eax,%edx,2)

 		ret

 

-/*

- * void v_putpixel24 (void *buffer, int offset, int color);

+/* Desc: void v_putpixel24 (unsigned int offset, int color);

+ *

+ * In  : offset within buffer, color

+ * Out : -

+ *

+ * Note: uses current draw buffer

  */

-		.balign	4

+		.p2align 5,,31

 		.global	_v_putpixel24

 _v_putpixel24:

+		movl	4(%esp), %eax

 		movl	8(%esp), %edx

-		leal	(%edx, %edx, 2), %edx

-		movl	12(%esp), %eax

-		addl	4(%esp), %edx

-		movw	%ax, (%edx)

-		shrl	$16, %eax

-		movb	%al, 2(%edx)

+		movl	_vl_current_draw_buffer, %ecx

+		leal	(%eax,%eax,2), %eax

+		movw	%dx, (%ecx,%eax)

+		shrl	$16, %edx

+		movb	%dl, 2(%ecx,%eax)

 		ret

 

-/*

- * void v_putpixel32 (void *buffer, int offset, int color);

+/* Desc: void v_putpixel32 (unsigned int offset, int color);

+ *

+ * In  : offset within buffer, color

+ * Out : -

+ *

+ * Note: uses current draw buffer

  */

-		.balign	4

+		.p2align 5,,31

 		.global	_v_putpixel32

 _v_putpixel32:

-		movl	8(%esp), %edx

-		shll	$2, %edx

-		movl	12(%esp), %eax

-		addl	4(%esp), %edx

-		movl	%eax, (%edx)

+		movl	8(%esp), %ecx

+		movl	4(%esp), %edx

+		movl	_vl_current_draw_buffer, %eax

+		movl	%ecx, (%eax,%edx,4)

 		ret