Source code of Windows XP (NT5)
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 

1090 lines
35 KiB

/******************************Module*Header*******************************\
* Module Name: blt.c
*
* Contains the low-level in/out blt functions.
*
* Hopefully, if you're basing your display driver on this code, to
* support all of DrvBitBlt and DrvCopyBits, you'll only have to implement
* the following routines. You shouldn't have to modify much in
* 'bitblt.c'. I've tried to make these routines as few, modular, simple,
* and efficient as I could, while still accelerating as many calls as
* possible that would be cost-effective in terms of performance wins
* versus size and effort.
*
* Note: In the following, 'relative' coordinates refers to coordinates
* that haven't yet had the offscreen bitmap (DFB) offset applied.
* 'Absolute' coordinates have had the offset applied. For example,
* we may be told to blt to (1, 1) of the bitmap, but the bitmap may
* be sitting in offscreen memory starting at coordinate (0, 768) --
* (1, 1) would be the 'relative' start coordinate, and (1, 769)
* would be the 'absolute' start coordinate'.
*
* Copyright (c) 1992-1995 Microsoft Corporation
*
\**************************************************************************/
#include "precomp.h"
/******************************Public*Table********************************\
* BYTE gaulP9000OpaqueFromRop2[]
*
* Convert an opaque Rop2 to a P9000 minterm.
*
\**************************************************************************/
#define P9000_P ((P9000_S & P9000_F) | (~P9000_S & P9000_B))
#define P9000_DSo (P9000_S | P9000_D)
#define P9000_DSna (~P9000_S & P9000_D)
ULONG gaulP9000OpaqueFromRop2[] = {
(0 ) & 0xFFFF, // 0 -- 0
(~(P9000_P | P9000_D)) & 0xFFFF, // 1 -- DPon
(~P9000_P & P9000_D ) & 0xFFFF, // 2 -- DPna
(~P9000_P ) & 0xFFFF, // 3 -- Pn
(~P9000_D & P9000_P ) & 0xFFFF, // 4 -- PDna
(~P9000_D ) & 0xFFFF, // 5 -- Dn
(P9000_D ^ P9000_P ) & 0xFFFF, // 6 -- DPx
(~(P9000_P & P9000_D)) & 0xFFFF, // 7 -- DPan
(P9000_P & P9000_D ) & 0xFFFF, // 8 -- DPa
(~(P9000_P ^ P9000_D)) & 0xFFFF, // 9 -- DPxn
(P9000_D ) & 0xFFFF, // 10 -- D
(~P9000_P | P9000_D ) & 0xFFFF, // 11 -- DPno
(P9000_P ) & 0xFFFF, // 12 -- P
(~P9000_D | P9000_P ) & 0xFFFF, // 13 -- PDno
(P9000_P | P9000_D ) & 0xFFFF, // 14 -- DPo
(~0 ) & 0xFFFF, // 15 -- 1
};
/******************************Public*Table********************************\
* BYTE gaulP9000TransparentFromRop2[]
*
* Convert a transparent Rop2 to a P9000 minterm.
*
\**************************************************************************/
ULONG gaulP9000TransparentFromRop2[] = {
(((0 ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 0 -- 0
(((~(P9000_P | P9000_D)) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 1 -- DPon
(((~P9000_P & P9000_D ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 2 -- DPna
(((~P9000_P ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 3 -- Pn
(((~P9000_D & P9000_P ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 4 -- PDna
(((~P9000_D ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 5 -- Dn
(((P9000_D ^ P9000_P ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 6 -- DPx
(((~(P9000_P & P9000_D)) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 7 -- DPan
(((P9000_P & P9000_D ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 8 -- DPa
(((~(P9000_P ^ P9000_D)) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 9 -- DPxn
(((P9000_D ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 10 -- D
(((~P9000_P | P9000_D ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 11 -- DPno
(((P9000_P ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 12 -- P
(((~P9000_D | P9000_P ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 13 -- PDno
(((P9000_P | P9000_D ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 14 -- DPo
(((~0 ) & P9000_DSo) | P9000_DSna) & 0xFFFF, // 15 -- 1
};
/******************************Public*Routine******************************\
* VOID vFillSolid
*
* Fills a list of rectangles with a solid colour.
*
\**************************************************************************/
VOID vFillSolid( // Type FNFILL
PDEV* ppdev,
LONG c, // Can't be zero
RECTL* prcl, // List of rectangles to be filled, in relative
// coordinates
ULONG ulHwMix, // Hardware mix mode
RBRUSH_COLOR rbc, // Drawing colour is rbc.iSolidColor
POINTL* pptlBrush) // Not used
{
BYTE* pjBase;
ASSERTDD(c > 0, "Can't handle zero rectangles");
pjBase = ppdev->pjBase;
CP_METARECT(ppdev, pjBase, prcl->left, prcl->top);
CP_METARECT(ppdev, pjBase, prcl->right, prcl->bottom);
CP_WAIT(ppdev, pjBase);
if (P9000(ppdev))
{
CP_BACKGROUND(ppdev, pjBase, rbc.iSolidColor);
CP_RASTER(ppdev, pjBase, ulHwMix);
}
else
{
CP_COLOR0(ppdev, pjBase, rbc.iSolidColor);
CP_RASTER(ppdev, pjBase, ulHwMix & 0xff);
}
CP_START_QUAD(ppdev, pjBase);
while (prcl++, --c)
{
CP_METARECT(ppdev, pjBase, prcl->left, prcl->top);
CP_METARECT(ppdev, pjBase, prcl->right, prcl->bottom);
CP_START_QUAD_WAIT(ppdev, pjBase);
}
}
/******************************Public*Routine******************************\
* VOID vSlowPatRealize
*
* This routine transfers an 8x8 pattern to off-screen display memory, and
* duplicates it to make a 64x64 cached realization which is then used by
* vFillPatSlow as the basic building block for doing 'slow' pattern output
* via repeated screen-to-screen blts.
*
\**************************************************************************/
VOID vSlowPatRealize(
PDEV* ppdev,
RBRUSH* prb) // Points to brush realization structure
{
BYTE* pjBase;
LONG cjPel;
BRUSHENTRY* pbe;
LONG iBrushCache;
LONG x;
LONG y;
ULONG* pulSrc;
LONG i;
ASSERTDD(!(prb->fl & (RBRUSH_2COLOR | RBRUSH_4COLOR)),
"Shouldn't realize hardware brushes");
pbe = prb->apbe[IBOARD(ppdev)];
if ((pbe == NULL) || (pbe->prbVerify != prb))
{
// We have to allocate a new off-screen cache brush entry for
// the brush:
iBrushCache = ppdev->iBrushCache;
pbe = &ppdev->abe[iBrushCache];
iBrushCache++;
if (iBrushCache >= ppdev->cBrushCache)
iBrushCache = 0;
ppdev->iBrushCache = iBrushCache;
// Update our links:
pbe->prbVerify = prb;
prb->apbe[IBOARD(ppdev)] = pbe;
}
pjBase = ppdev->pjBase;
cjPel = ppdev->cjPel;
// Load some pointer variables onto the stack, so that we don't have
// to keep dereferencing their pointers:
x = pbe->x;
y = pbe->y;
CP_ABS_XY0(ppdev, pjBase, x * cjPel, y);
CP_ABS_XY1(ppdev, pjBase, x * cjPel, y);
CP_ABS_XY2(ppdev, pjBase, (x + 8) * cjPel, y);
CP_ABS_Y3(ppdev, pjBase, 1);
pulSrc = (ULONG*) &prb->aulPattern[0];
CP_WAIT(ppdev, pjBase);
CP_RASTER(ppdev, pjBase, P9000(ppdev) ? P9000_S : P9100_S);
for (i = 4 * cjPel; i != 0; i--)
{
CP_PIXEL8(ppdev, pjBase, *(pulSrc));
CP_PIXEL8(ppdev, pjBase, *(pulSrc + 1));
CP_PIXEL8(ppdev, pjBase, *(pulSrc + 2));
CP_PIXEL8(ppdev, pjBase, *(pulSrc + 3));
pulSrc += 4;
}
// ÚÄÂÄÂÄÂÄÄÄÂÄÄÄÄÄÄÄ¿
// ³0³1³2³3 ³4 ³ We now have an 8x8 colour-expanded copy of
// ÃÄÁÄÁÄÁÄÄÄÁÄÄÄÄÄÄÄ´ the pattern sitting in off-screen memory,
// ³5 ³ represented here by square '0'.
// ÃÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ´
// ³6 ³ We're now going to expand the pattern to
// ÃÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ´ 72x72 by repeatedly copying larger rectangles
// ³7 ³ in the indicated order.
// ³ ³
// ³ ³
// ÃÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ´
// ³8 ³
// ³ ³
// ³ ³
// ³ ³
// ³ ³
// ÀÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÙ
// Copy '1':
CP_ABS_XY1(ppdev, pjBase, x + 7, y + 7);
CP_ABS_XY3(ppdev, pjBase, x + 15, y + 7);
CP_WAIT(ppdev, pjBase);
CP_START_BLT(ppdev, pjBase);
// Copy '2':
CP_ABS_X2(ppdev, pjBase, x + 16);
CP_ABS_X3(ppdev, pjBase, x + 23);
CP_START_BLT_WAIT(ppdev, pjBase);
// Copy '3':
CP_ABS_X1(ppdev, pjBase, x + 15);
CP_ABS_X2(ppdev, pjBase, x + 24);
CP_ABS_X3(ppdev, pjBase, x + 39);
CP_START_BLT_WAIT(ppdev, pjBase);
// Copy '4':
CP_ABS_X1(ppdev, pjBase, x + 31);
CP_ABS_X2(ppdev, pjBase, x + 40);
CP_ABS_X3(ppdev, pjBase, x + 71);
CP_START_BLT_WAIT(ppdev, pjBase);
// Copy '5':
CP_ABS_XY1(ppdev, pjBase, x + 71, y + 7);
CP_ABS_XY2(ppdev, pjBase, x, y + 8);
CP_ABS_XY3(ppdev, pjBase, x + 71, y + 15);
CP_START_BLT_WAIT(ppdev, pjBase);
// Copy '6':
CP_ABS_Y2(ppdev, pjBase, y + 16);
CP_ABS_Y3(ppdev, pjBase, y + 23);
CP_START_BLT_WAIT(ppdev, pjBase);
// Copy '7':
CP_ABS_Y1(ppdev, pjBase, y + 15);
CP_ABS_Y2(ppdev, pjBase, y + 24);
CP_ABS_Y3(ppdev, pjBase, y + 39);
CP_START_BLT_WAIT(ppdev, pjBase);
// Copy '8':
CP_ABS_Y1(ppdev, pjBase, y + 31);
CP_ABS_Y2(ppdev, pjBase, y + 40);
CP_ABS_Y3(ppdev, pjBase, y + 71);
CP_START_BLT_WAIT(ppdev, pjBase);
}
/******************************Public*Routine******************************\
* VOID vFillSlowPat
*
\**************************************************************************/
VOID vFillSlowPat( // Type FNFILL
PDEV* ppdev,
LONG c, // Can't be zero
RECTL* prcl, // List of rectangles to be filled, in relative
// coordinates
ULONG ulHwMix, // Hardware mix mode
RBRUSH_COLOR rbc, // rbc.prb points to brush realization structure
POINTL* pptlBrush) // Pattern alignment
{
BYTE* pjBase;
BOOL bExponential;
ULONG ulRaster;
LONG xBrush;
LONG yBrush;
LONG xSrc;
LONG ySrc;
LONG x;
LONG y;
LONG xFrom;
LONG yFrom;
LONG cxToGo;
LONG cyToGo;
LONG cxThis;
LONG cyThis;
LONG xOrg;
LONG yOrg;
LONG cyOriginal;
LONG yOriginal;
LONG xOriginal;
BRUSHENTRY* pbe; // Pointer to brush entry data, which is used
// for keeping track of the location and status
// of the pattern bits cached in off-screen
// memory
if (rbc.prb->apbe[IBOARD(ppdev)]->prbVerify != rbc.prb)
{
vSlowPatRealize(ppdev, rbc.prb);
}
pjBase = ppdev->pjBase;
// We special case PATCOPY mixes because we can implement
// an exponential fill: every blt will double the size of
// the current rectangle by using the portion of the pattern
// that has already been done for this rectangle as the source.
//
// Note that there's no point in also checking for BLACK
// or WHITE because those will be taken care of by the
// solid fill routines, and I can't be bothered to check for
// NOTCOPYPEN:
bExponential = (ulHwMix == 0xf0f0);
// Convert the rop from a Rop3 between P and D to the corresponding
// Rop3 between S and D:
ulRaster = (ulHwMix & 0x3C) >> 2;
ulRaster |= (ulRaster << 4);
if (P9000(ppdev))
{
// Make the Rop3 into a true P9000 minterm:
ulRaster |= (ulRaster << 8);
}
// Note that since we do our brush alignment calculations in
// relative coordinates, we should keep the brush origin in
// relative coordinates as well:
xOrg = pptlBrush->x;
yOrg = pptlBrush->y;
pbe = rbc.prb->apbe[IBOARD(ppdev)];
xBrush = pbe->x;
yBrush = pbe->y;
do {
x = prcl->left;
y = prcl->top;
xSrc = xBrush + ((x - xOrg) & 7);
ySrc = yBrush + ((y - yOrg) & 7);
cxToGo = prcl->right - x;
cyToGo = prcl->bottom - y;
if ((cxToGo <= SLOW_BRUSH_DIMENSION) &&
(cyToGo <= SLOW_BRUSH_DIMENSION))
{
CP_ABS_XY0(ppdev, pjBase, xSrc, ySrc);
CP_ABS_XY1(ppdev, pjBase, xSrc + cxToGo - 1, ySrc + cyToGo - 1);
CP_XY2(ppdev, pjBase, x, y);
CP_XY3(ppdev, pjBase, x + cxToGo - 1, y + cyToGo - 1);
CP_WAIT(ppdev, pjBase);
CP_RASTER(ppdev, pjBase, ulRaster);
CP_START_BLT(ppdev, pjBase);
}
else if (bExponential)
{
cyThis = SLOW_BRUSH_DIMENSION;
cyToGo -= cyThis;
if (cyToGo < 0)
cyThis += cyToGo;
cxThis = SLOW_BRUSH_DIMENSION;
cxToGo -= cxThis;
if (cxToGo < 0)
cxThis += cxToGo;
CP_ABS_XY0(ppdev, pjBase, xSrc, ySrc);
CP_ABS_XY1(ppdev, pjBase, xSrc + cxThis - 1, ySrc + cyThis - 1);
CP_XY2(ppdev, pjBase, x, y);
CP_XY3(ppdev, pjBase, x + cxThis - 1, y + cyThis - 1);
CP_WAIT(ppdev, pjBase);
CP_RASTER(ppdev, pjBase, ulRaster);
CP_START_BLT(ppdev, pjBase);
CP_XY0(ppdev, pjBase, x, y);
xOriginal = x;
x += cxThis;
y += cyThis;
while (cxToGo > 0)
{
// First, expand out to the right, doubling our size
// each time:
xFrom = x;
cxToGo -= cxThis;
if (cxToGo < 0)
{
cxThis += cxToGo;
xFrom += cxToGo;
}
CP_XY1(ppdev, pjBase, xFrom - 1, y - 1);
CP_X2(ppdev, pjBase, x);
CP_X3(ppdev, pjBase, x + cxThis - 1);
CP_START_BLT_WAIT(ppdev, pjBase);
x += cxThis;
cxThis *= 2;
}
while (cyToGo > 0)
{
// Now do the same thing vertically:
yFrom = y;
cyToGo -= cyThis;
if (cyToGo < 0)
{
cyThis += cyToGo;
yFrom += cyToGo;
}
CP_XY1(ppdev, pjBase, x - 1, yFrom - 1);
CP_XY2(ppdev, pjBase, xOriginal, y);
CP_Y3(ppdev, pjBase, y + cyThis - 1);
CP_START_BLT_WAIT(ppdev, pjBase);
y += cyThis;
cyThis *= 2;
}
}
else
{
// We handle arbitrary mixes simply by repeatedly tiling
// our cached pattern over the entire rectangle:
CP_ABS_XY0(ppdev, pjBase, xSrc, ySrc);
CP_WAIT(ppdev, pjBase);
CP_RASTER(ppdev, pjBase, ulRaster);
cyOriginal = cyToGo; // Have to remember for later...
yOriginal = y;
do {
cxThis = SLOW_BRUSH_DIMENSION;
cxToGo -= cxThis;
if (cxToGo < 0)
cxThis += cxToGo;
cyToGo = cyOriginal; // Have to reset for each new column
y = yOriginal;
do {
cyThis = SLOW_BRUSH_DIMENSION;
cyToGo -= cyThis;
if (cyToGo < 0)
cyThis += cyToGo;
CP_ABS_XY1(ppdev, pjBase, xSrc + cxThis - 1,
ySrc + cyThis - 1);
CP_XY2(ppdev, pjBase, x, y);
CP_XY3(ppdev, pjBase, x + cxThis - 1,
y + cyThis - 1);
CP_START_BLT_WAIT(ppdev, pjBase);
y += cyThis;
} while (cyToGo > 0);
x += cxThis; // Get ready for next column
} while (cxToGo > 0);
}
prcl++;
} while (--c != 0);
}
/******************************Public*Routine******************************\
* VOID vFillPat
*
\**************************************************************************/
VOID vFillPat( // Type FNFILL
PDEV* ppdev,
LONG c, // Can't be zero
RECTL* prcl, // List of rectangles to be filled, in relative
// coordinates
ULONG ulHwMix, // Hardware mix mode
RBRUSH_COLOR rbc, // rbc.prb points to brush realization structure
POINTL* pptlBrush) // Pattern alignment
{
BYTE* pjBase;
LONG i;
ULONG* pulPattern;
ULONG ulPattern;
ASSERTDD(((ulHwMix >> 8) == (ulHwMix & 0xff)) ||
((ulHwMix & 0xff00) == 0xaa00),
"This routine handles only opaque or transparent mixes");
if (rbc.prb->fl & (RBRUSH_2COLOR | RBRUSH_4COLOR))
{
// Hardware pattern
pjBase = ppdev->pjBase;
CP_METARECT(ppdev, pjBase, prcl->left, prcl->top);
CP_METARECT(ppdev, pjBase, prcl->right, prcl->bottom);
CP_WAIT(ppdev, pjBase);
if (P9000(ppdev))
{
CP_PATTERN_ORGX(ppdev, pjBase, ppdev->xOffset + pptlBrush->x);
CP_PATTERN_ORGY(ppdev, pjBase, ppdev->yOffset + pptlBrush->y);
CP_BACKGROUND(ppdev, pjBase, rbc.prb->ulColor[0]);
CP_FOREGROUND(ppdev, pjBase, rbc.prb->ulColor[1]);
pulPattern = &rbc.prb->aulPattern[0];
for (i = 0; i < 4; i++)
{
ulPattern = *pulPattern++;
CP_PATTERN(ppdev, pjBase, i, ulPattern);
CP_PATTERN(ppdev, pjBase, i + 4, ulPattern);
}
if (((ulHwMix >> 8) & 0xff) == (ulHwMix & 0xff))
{
ulHwMix = gaulP9000OpaqueFromRop2[(ulHwMix & 0x3C) >> 2];
CP_RASTER(ppdev, pjBase, ulHwMix | P9000_ENABLE_PATTERN);
}
else
{
ulHwMix = gaulP9000TransparentFromRop2[(ulHwMix & 0x3C) >> 2];
CP_RASTER(ppdev, pjBase, ulHwMix | P9000_ENABLE_PATTERN);
}
}
else
{
CP_PATTERN_ORGX(ppdev, pjBase, -(ppdev->xOffset + pptlBrush->x));
CP_PATTERN_ORGY(ppdev, pjBase, -(ppdev->yOffset + pptlBrush->y));
CP_COLOR0_FAST(ppdev, pjBase, rbc.prb->ulColor[0]);
CP_COLOR1_FAST(ppdev, pjBase, rbc.prb->ulColor[1]);
CP_PATTERN(ppdev, pjBase, 0, rbc.prb->aulPattern[0]);
CP_PATTERN(ppdev, pjBase, 1, rbc.prb->aulPattern[1]);
CP_PATTERN(ppdev, pjBase, 2, rbc.prb->aulPattern[2]);
CP_PATTERN(ppdev, pjBase, 3, rbc.prb->aulPattern[3]);
if (rbc.prb->fl & RBRUSH_2COLOR)
{
if (((ulHwMix >> 8) & 0xff) == (ulHwMix & 0xff))
{
CP_RASTER(ppdev, pjBase, (ulHwMix & 0xff)
| P9100_ENABLE_PATTERN);
}
else
{
CP_RASTER(ppdev, pjBase, (ulHwMix & 0xff)
| P9100_ENABLE_PATTERN | P9100_TRANSPARENT_PATTERN);
}
}
else
{
CP_COLOR2_FAST(ppdev, pjBase, rbc.prb->ulColor[2]);
CP_COLOR3_FAST(ppdev, pjBase, rbc.prb->ulColor[3]);
CP_RASTER(ppdev, pjBase, (ulHwMix & 0xff)
| P9100_ENABLE_PATTERN | P9100_FOUR_COLOR_PATTERN);
}
}
CP_START_QUAD(ppdev, pjBase);
while (prcl++, --c)
{
CP_METARECT(ppdev, pjBase, prcl->left, prcl->top);
CP_METARECT(ppdev, pjBase, prcl->right, prcl->bottom);
CP_START_QUAD_WAIT(ppdev, pjBase);
}
}
else
{
vFillSlowPat(ppdev, c, prcl, ulHwMix, rbc, pptlBrush);
}
}
/******************************Public*Routine******************************\
* VOID vXfer1bpp
*
* This routine colour expands a monochrome bitmap.
*
\**************************************************************************/
VOID vXfer1bpp( // Type FNXFER
PDEV* ppdev,
LONG c, // Count of rectangles, can't be zero
RECTL* prcl, // List of destination rectangles, in relative
// coordinates
ULONG ulHwMix, // Foreground and background hardware mix
SURFOBJ* psoSrc, // Source surface
POINTL* pptlSrc, // Original unclipped source point
RECTL* prclDst, // Original unclipped destination rectangle
XLATEOBJ* pxlo) // Translate that provides colour-expansion information
{
BYTE* pjBase;
LONG dx;
LONG dy;
LONG lSrcDelta;
BYTE* pjSrcScan0;
ULONG* pulXlate;
LONG xLeft;
LONG xRight;
LONG yTop;
LONG cyScan;
LONG xBias;
LONG culScan;
LONG lSrcSkip;
ULONG* pulSrc;
LONG cRem;
LONG i;
ASSERTDD(((ulHwMix >> 8) & 0xff) == (ulHwMix & 0xff),
"Expected only an opaquing rop");
pjBase = ppdev->pjBase;
dx = pptlSrc->x - prclDst->left;
dy = pptlSrc->y - prclDst->top;
lSrcDelta = psoSrc->lDelta;
pjSrcScan0 = psoSrc->pvScan0;
do {
xLeft = prcl->left;
xRight = prcl->right;
yTop = prcl->top;
xBias = (xLeft + dx) & 31;
xLeft -= xBias;
CP_XY1(ppdev, pjBase, xLeft, yTop);
CP_X0(ppdev, pjBase, xLeft);
CP_X2(ppdev, pjBase, xRight);
CP_ABS_Y3(ppdev, pjBase, 1);
culScan = ((xRight - xLeft) >> 5);
cRem = ((xRight - xLeft) & 31) - 1;
if (cRem < 0)
{
culScan--;
cRem = 31;
}
pulSrc = (ULONG*) (pjSrcScan0 + (yTop + dy) * lSrcDelta
+ ((xLeft + dx) >> 3));
lSrcSkip = lSrcDelta - (culScan << 2);
cyScan = (prcl->bottom - yTop);
CP_WAIT(ppdev, pjBase);
CP_WLEFT(ppdev, pjBase, prcl->left);
// The following three accelerator states are invariant to this
// loop, but we set them each time anyway because we expect
// usually to have only to do one iteration of this loop, and this
// state must to be set after doing a CP_WAIT, which we want to
// delay until we get as much processing done as possible, for
// maximum overlap.
pulXlate = pxlo->pulXlate;
if (P9000(ppdev))
{
CP_BACKGROUND(ppdev, pjBase, pulXlate[0]);
CP_FOREGROUND(ppdev, pjBase, pulXlate[1]);
CP_RASTER(ppdev, pjBase, gaulP9000OpaqueFromRop2[ulHwMix & 0xf]);
}
else
{
CP_COLOR0(ppdev, pjBase, pulXlate[0]);
CP_COLOR1(ppdev, pjBase, pulXlate[1]);
CP_RASTER(ppdev, pjBase, ulHwMix & 0xff);
}
CP_START_PIXEL1(ppdev, pjBase);
do {
for (i = culScan; i != 0; i--)
{
CP_PIXEL1(ppdev, pjBase, *pulSrc);
pulSrc++;
}
CP_PIXEL1_REM(ppdev, pjBase, cRem, *pulSrc);
pulSrc = (ULONG*) ((BYTE*) pulSrc + lSrcSkip);
} while (--cyScan != 0);
CP_END_PIXEL1(ppdev, pjBase);
} while (prcl++, --c);
// Don't forget to reset the clip register:
CP_WAIT(ppdev, pjBase);
CP_ABS_WMIN(ppdev, pjBase, 0, 0);
}
/******************************Public*Routine******************************\
* VOID vXfer4bpp
*
* Does a 4bpp transfer from a bitmap to the screen.
*
* NOTE: The screen must be 8bpp for this function to be called!
*
* The reason we implement this is that a lot of resources are kept as 4bpp,
* and used to initialize DFBs, some of which we of course keep off-screen.
*
\**************************************************************************/
VOID vXfer4bpp( // Type FNXFER
PDEV* ppdev,
LONG c, // Count of rectangles, can't be zero
RECTL* prcl, // List of destination rectangles, in relative
// coordinates
ULONG ulHwMix, // Hardware mix
SURFOBJ* psoSrc, // Source surface
POINTL* pptlSrc, // Original unclipped source point
RECTL* prclDst, // Original unclipped destination rectangle
XLATEOBJ* pxlo) // Translate that provides colour-expansion information
{
BYTE* pjBase;
LONG dx;
LONG dy;
LONG lSrcDelta;
BYTE* pjSrcScan0;
ULONG* pulXlate;
LONG xLeft;
LONG xRight;
LONG yTop;
LONG cyScan;
LONG xBias;
LONG cwSrc;
LONG lSrcSkip;
LONG cw;
BYTE* pjSrc;
BYTE jSrc;
ULONG ul;
ASSERTDD(ppdev->cjPel == 1, "This function assumes 8bpp");
pjBase = ppdev->pjBase;
dx = pptlSrc->x - prclDst->left;
dy = pptlSrc->y - prclDst->top;
lSrcDelta = psoSrc->lDelta;
pjSrcScan0 = psoSrc->pvScan0;
pulXlate = pxlo->pulXlate;
do {
xLeft = prcl->left;
xRight = prcl->right;
yTop = prcl->top;
// We compute 'xBias' in order to word-align the source pointer.
// This way, since we're processing a word of the source at a
// time, we're guaranteed not to read even a byte past the end of
// the bitmap.
xBias = ((xLeft + dx) & 3);
xLeft -= xBias;
CP_XY1(ppdev, pjBase, xLeft, yTop);
CP_X0(ppdev, pjBase, xLeft);
CP_X2(ppdev, pjBase, xRight);
CP_ABS_Y3(ppdev, pjBase, 1);
// Every word in the source is one dword in the destination:
cwSrc = ((xRight - xLeft) + 3) >> 2;
lSrcSkip = lSrcDelta - (cwSrc << 1);
pjSrc = pjSrcScan0 + (yTop + dy) * lSrcDelta + ((xLeft + dx) >> 1);
cyScan = prcl->bottom - yTop;
ASSERTDD(((ULONG_PTR) pjSrc & 1) == 0, "Source should be word aligned");
// Yes, we're setting the raster every time in the loop. But we
// have to wait for not-busy before setting the raster, and the
// chances are that we'll only have one rectangle to blt, so
// we do this here:
CP_WAIT(ppdev, pjBase);
CP_WLEFT(ppdev, pjBase, prcl->left);
CP_RASTER(ppdev, pjBase, P9000(ppdev) ? ulHwMix : (ulHwMix & 0xff));
CP_START_PIXEL8(ppdev, pjBase);
do {
cw = cwSrc;
do {
jSrc = *(pjSrc + 1);
ul = pulXlate[jSrc & 0xf];
ul <<= 8;
ul |= pulXlate[jSrc >> 4];
jSrc = *(pjSrc);
ul <<= 8;
ul |= pulXlate[jSrc & 0xf];
ul <<= 8;
ul |= pulXlate[jSrc >> 4];
CP_PIXEL8(ppdev, pjBase, ul);
pjSrc += 2;
} while (--cw != 0);
pjSrc += lSrcSkip;
} while (--cyScan != 0);
CP_END_PIXEL8(ppdev, pjBase);
} while (prcl++, --c);
// Don't forget to reset the clip register:
CP_WAIT(ppdev, pjBase);
CP_ABS_WMIN(ppdev, pjBase, 0, 0);
}
/******************************Public*Routine******************************\
* VOID vXferNative
*
* Transfers a bitmap that is the same colour depth as the display to
* the screen via the data transfer register, with no translation.
*
\**************************************************************************/
VOID vXferNative( // Type FNXFER
PDEV* ppdev,
LONG c, // Count of rectangles, can't be zero
RECTL* prcl, // Array of relative coordinates destination rectangles
ULONG ulHwMix, // Hardware mix
SURFOBJ* psoSrc, // Source surface
POINTL* pptlSrc, // Original unclipped source point
RECTL* prclDst, // Original unclipped destination rectangle
XLATEOBJ* pxlo) // Not used
{
BYTE* pjBase;
LONG cjPel;
LONG xOffset;
LONG yOffset;
LONG dx;
LONG dy;
LONG lSrcDelta;
BYTE* pjSrcScan0;
LONG xLeft;
LONG xRight;
LONG yTop;
LONG cyScan;
LONG xBias;
LONG culScan;
LONG lSrcSkip;
LONG cu;
ULONG* pulSrc;
pjBase = ppdev->pjBase;
cjPel = ppdev->cjPel;
xOffset = ppdev->xOffset;
yOffset = ppdev->yOffset;
dx = pptlSrc->x - prclDst->left;
dy = pptlSrc->y - prclDst->top;
lSrcDelta = psoSrc->lDelta;
pjSrcScan0 = psoSrc->pvScan0;
do {
xLeft = prcl->left;
xRight = prcl->right;
yTop = prcl->top;
// We compute 'xBias' in order to dword-align the source pointer.
// This way, we don't have to do unaligned reads of the source,
// and we're guaranteed not to read even a byte past the end of
// the bitmap.
//
// Note that this bias works at 24bpp, too:
xBias = ((xLeft + dx) & 3);
xLeft -= xBias;
CP_ABS_XY1(ppdev, pjBase, (xLeft + xOffset) * cjPel, yTop + yOffset);
CP_ABS_X0(ppdev, pjBase, (xLeft + xOffset) * cjPel);
CP_ABS_X2(ppdev, pjBase, (xRight + xOffset) * cjPel);
CP_ABS_Y3(ppdev, pjBase, 1);
culScan = ((xRight - xLeft) * cjPel + 3) >> 2;
lSrcSkip = lSrcDelta - (culScan << 2);
pulSrc = (ULONG*) (pjSrcScan0 + (yTop + dy) * lSrcDelta
+ (xLeft + dx) * cjPel);
cyScan = prcl->bottom - yTop;
ASSERTDD(((ULONG_PTR)pulSrc & 3) == 0, "Source should be dword aligned");
// Yes, we're setting the raster every time in the loop. But we
// have to wait for not-busy before setting the raster, and the
// chances are that we'll only have one rectangle to blt, so
// we do this here:
CP_WAIT(ppdev, pjBase);
CP_ABS_WLEFT(ppdev, pjBase, prcl->left + xOffset);
CP_RASTER(ppdev, pjBase, P9000(ppdev) ? ulHwMix : (ulHwMix & 0xff));
CP_START_PIXEL8(ppdev, pjBase);
do {
cu = culScan;
do {
CP_PIXEL8(ppdev, pjBase, *pulSrc);
pulSrc++;
} while (--cu != 0);
pulSrc = (ULONG*) ((BYTE*) pulSrc + lSrcSkip);
} while (--cyScan != 0);
CP_END_PIXEL8(ppdev, pjBase);
} while (prcl++, --c);
// Don't forget to reset the clip register:
CP_WAIT(ppdev, pjBase);
CP_ABS_WMIN(ppdev, pjBase, 0, 0);
}
/******************************Public*Routine******************************\
* VOID vCopyBlt
*
* Does a screen-to-screen blt of a list of rectangles.
*
* Note: We may call this function with weird ROPs to do colour-expansion
* from off-screen monochrome bitmaps.
*
\**************************************************************************/
VOID vCopyBlt( // Type FNCOPY
PDEV* ppdev,
LONG c, // Can't be zero
RECTL* prcl, // Array of relative coordinates destination rectangles
ULONG ulHwMix, // Hardware mix
POINTL* pptlSrc, // Original unclipped source point
RECTL* prclDst) // Original unclipped destination rectangle
{
BYTE* pjBase;
LONG cjPel;
LONG xOffset;
LONG yOffset;
LONG dx;
LONG dy;
LONG xSrc;
LONG ySrc;
pjBase = ppdev->pjBase;
cjPel = ppdev->cjPel;
// Our DFB xOffset has to be scaled by the pixel size too:
xOffset = ppdev->xOffset;
yOffset = ppdev->yOffset;
dx = pptlSrc->x - prclDst->left;
dy = pptlSrc->y - prclDst->top;
xSrc = prcl->left + dx;
ySrc = prcl->top + dy;
CP_ABS_XY0(ppdev, pjBase, (xOffset + xSrc) * cjPel,
(yOffset + ySrc));
CP_ABS_XY1(ppdev, pjBase, (xOffset + xSrc + prcl->right - prcl->left) * cjPel - 1,
(yOffset + ySrc + prcl->bottom - prcl->top - 1));
CP_ABS_XY2(ppdev, pjBase, (xOffset + prcl->left) * cjPel,
(yOffset + prcl->top));
CP_ABS_XY3(ppdev, pjBase, (xOffset + prcl->right) * cjPel - 1,
(yOffset + prcl->bottom - 1));
CP_WAIT(ppdev, pjBase);
CP_RASTER(ppdev, pjBase, P9000(ppdev) ? ulHwMix : (ulHwMix & 0xff));
CP_START_BLT(ppdev, pjBase);
while (prcl++, --c)
{
xSrc = prcl->left + dx;
ySrc = prcl->top + dy;
CP_ABS_XY0(ppdev, pjBase, (xOffset + xSrc) * cjPel,
(yOffset + ySrc));
CP_ABS_XY1(ppdev, pjBase, (xOffset + xSrc + prcl->right - prcl->left) * cjPel - 1,
(yOffset + ySrc + prcl->bottom - prcl->top - 1));
CP_ABS_XY2(ppdev, pjBase, (xOffset + prcl->left) * cjPel,
(yOffset + prcl->top));
CP_ABS_XY3(ppdev, pjBase, (xOffset + prcl->right) * cjPel - 1,
(yOffset + prcl->bottom - 1));
CP_START_BLT_WAIT(ppdev, pjBase);
}
}
/******************************Public*Routine******************************\
* VOID vFillSolidP9000HighColor
*
* Fills a list of rectangles with a solid colour when running at 16bpp
* on the P9000, using the pattern hardware.
*
\**************************************************************************/
VOID vFillSolidP9000HighColor( // Type FNFILL
PDEV* ppdev,
LONG c, // Can't be zero
RECTL* prcl, // List of rectangles to be filled, in relative
// coordinates
ULONG ulHwMix, // Hardware mix mode
RBRUSH_COLOR rbc, // Drawing colour is rbc.iSolidColor
POINTL* pptlBrush) // Not used
{
BYTE* pjBase;
LONG xOffset;
LONG yOffset;
ASSERTDD(c > 0, "Can't handle zero rectangles");
ASSERTDD(P9000(ppdev), "Shouldn't need to be called on the P9100");
ASSERTDD(ppdev->iBitmapFormat == BMF_16BPP, "Only handle 16bpp for now");
pjBase = ppdev->pjBase;
// Our DFB xOffset has to be scaled by the pixel size too:
xOffset = 2 * ppdev->xOffset;
yOffset = ppdev->yOffset;
CP_ABS_METARECT(ppdev, pjBase, xOffset + 2 * prcl->left,
yOffset + prcl->top);
CP_ABS_METARECT(ppdev, pjBase, xOffset + 2 * prcl->right,
yOffset + prcl->bottom);
// We've already downloaded and set the pattern origin in
// vAssertModeBrushCache:
CP_WAIT(ppdev, pjBase);
CP_RASTER(ppdev, pjBase, P9000_ENABLE_PATTERN |
gaulP9000OpaqueFromRop2[(ulHwMix & 0x3C) >> 2]);
CP_FOREGROUND(ppdev, pjBase, rbc.iSolidColor);
CP_BACKGROUND(ppdev, pjBase, rbc.iSolidColor >> 8);
CP_START_QUAD(ppdev, pjBase);
while (prcl++, --c)
{
CP_ABS_METARECT(ppdev, pjBase, xOffset + 2 * prcl->left,
yOffset + prcl->top);
CP_ABS_METARECT(ppdev, pjBase, xOffset + 2 * prcl->right,
yOffset + prcl->bottom);
CP_START_QUAD_WAIT(ppdev, pjBase);
}
}