mirror of
https://github.com/holub/mame
synced 2025-04-22 08:22:15 +03:00
Further DMA simplfication in Kangaroo, moving it closer to a merge with Arabian.
This commit is contained in:
Aaron Giles
parent
c87f5ef4f8
commit
b3cd0a8e86
@ -24,8 +24,49 @@ static void blitter_execute(void);
|
||||
|
||||
VIDEO_START( kangaroo )
|
||||
{
|
||||
videoram = auto_malloc(256 * 256);
|
||||
state_save_register_global_pointer(videoram, 256 * 256);
|
||||
/* video RAM is accessed 32 bits at a time (two planes, 4bpp each, 4 pixels) */
|
||||
videoram32 = auto_malloc(256 * 64 * 4);
|
||||
state_save_register_global_pointer(videoram32, 256 * 64);
|
||||
}
|
||||
|
||||
|
||||
|
||||
/*************************************
|
||||
*
|
||||
* Video RAM accesses
|
||||
*
|
||||
*************************************/
|
||||
|
||||
static void videoram_write(UINT16 offset, UINT8 data, UINT8 mask)
|
||||
{
|
||||
UINT32 expdata, layermask;
|
||||
|
||||
/* data contains 4 2-bit values packed as DCBADCBA; expand these into 4 8-bit values */
|
||||
expdata = 0;
|
||||
if (data & 0x01) expdata |= 0x00000055;
|
||||
if (data & 0x10) expdata |= 0x000000aa;
|
||||
if (data & 0x02) expdata |= 0x00005500;
|
||||
if (data & 0x20) expdata |= 0x0000aa00;
|
||||
if (data & 0x04) expdata |= 0x00550000;
|
||||
if (data & 0x40) expdata |= 0x00aa0000;
|
||||
if (data & 0x08) expdata |= 0x55000000;
|
||||
if (data & 0x80) expdata |= 0xaa000000;
|
||||
|
||||
/* determine which layers are enabled */
|
||||
layermask = 0;
|
||||
if (mask & 0x08) layermask |= 0x30303030;
|
||||
if (mask & 0x04) layermask |= 0xc0c0c0c0;
|
||||
if (mask & 0x02) layermask |= 0x03030303;
|
||||
if (mask & 0x01) layermask |= 0x0c0c0c0c;
|
||||
|
||||
/* update layers */
|
||||
videoram32[offset] = (videoram32[offset] & ~layermask) | (expdata & layermask);
|
||||
}
|
||||
|
||||
|
||||
WRITE8_HANDLER( kangaroo_videoram_w )
|
||||
{
|
||||
videoram_write(offset, data, kangaroo_video_control[8]);
|
||||
}
|
||||
|
||||
|
||||
@ -62,100 +103,29 @@ WRITE8_HANDLER( kangaroo_video_control_w )
|
||||
|
||||
static void blitter_execute(void)
|
||||
{
|
||||
int src,dest;
|
||||
int x,y,width,height,old_bank_select,new_bank_select;
|
||||
UINT32 gfxhalfsize = memory_region_length(REGION_GFX1) / 2;
|
||||
const UINT8 *gfxbase = memory_region(REGION_GFX1);
|
||||
UINT16 src = kangaroo_video_control[0] + 256 * kangaroo_video_control[1];
|
||||
UINT16 dst = kangaroo_video_control[2] + 256 * kangaroo_video_control[3];
|
||||
UINT8 height = kangaroo_video_control[5];
|
||||
UINT8 width = kangaroo_video_control[4];
|
||||
UINT8 mask = kangaroo_video_control[8];
|
||||
int x, y;
|
||||
|
||||
/* during DMA operations, the top 2 bits are ORed together, as well as the bottom 2 bits */
|
||||
/* adjust the mask to account for this */
|
||||
if (mask & 0x0c) mask |= 0x0c;
|
||||
if (mask & 0x03) mask |= 0x03;
|
||||
|
||||
src = kangaroo_video_control[0] + 256 * kangaroo_video_control[1];
|
||||
dest = kangaroo_video_control[2] + 256 * kangaroo_video_control[3];
|
||||
|
||||
width = kangaroo_video_control[5];
|
||||
height = kangaroo_video_control[4];
|
||||
|
||||
old_bank_select = new_bank_select = kangaroo_video_control[8];
|
||||
|
||||
if (new_bank_select & 0x0c) new_bank_select |= 0x0c;
|
||||
if (new_bank_select & 0x03) new_bank_select |= 0x03;
|
||||
|
||||
kangaroo_video_control_w(8, new_bank_select & 0x05);
|
||||
|
||||
for (x = 0; x <= width; x++)
|
||||
{
|
||||
for (y = 0; y <= height; y++)
|
||||
program_write_byte(dest + y, program_read_byte(src++));
|
||||
|
||||
dest += 256;
|
||||
}
|
||||
|
||||
src = kangaroo_video_control[0] + 256 * kangaroo_video_control[1];
|
||||
dest = kangaroo_video_control[2] + 256 * kangaroo_video_control[3];
|
||||
|
||||
kangaroo_video_control_w(8, new_bank_select & 0x0a);
|
||||
|
||||
for (x = 0; x <= width; x++)
|
||||
{
|
||||
for (y = 0; y <= height; y++)
|
||||
program_write_byte(dest + y, program_read_byte(src++));
|
||||
|
||||
dest += 256;
|
||||
}
|
||||
|
||||
kangaroo_video_control_w(8, old_bank_select);
|
||||
}
|
||||
|
||||
|
||||
|
||||
/*************************************
|
||||
*
|
||||
* Video RAM writes
|
||||
*
|
||||
*************************************/
|
||||
|
||||
WRITE8_HANDLER( kangaroo_videoram_w )
|
||||
{
|
||||
int sx, sy, offs;
|
||||
|
||||
sx = (offset / 256) * 4;
|
||||
sy = offset % 256;
|
||||
offs = sy * 256 + sx;
|
||||
|
||||
/* rearrange the bits to a more convenient pattern (from DCBADCBA to DDCCBBAA) */
|
||||
data = BITSWAP8(data, 7,3,6,2,5,1,4,0);
|
||||
|
||||
/* B layer, green & blue bits */
|
||||
if (kangaroo_video_control[8] & 0x08)
|
||||
{
|
||||
videoram[offs ] = (videoram[offs ] & 0xcf) | (((data >> 0) & 3) << 4);
|
||||
videoram[offs+1] = (videoram[offs+1] & 0xcf) | (((data >> 2) & 3) << 4);
|
||||
videoram[offs+2] = (videoram[offs+2] & 0xcf) | (((data >> 4) & 3) << 4);
|
||||
videoram[offs+3] = (videoram[offs+3] & 0xcf) | (((data >> 6) & 3) << 4);
|
||||
}
|
||||
|
||||
/* B layer, Z & red bits */
|
||||
if (kangaroo_video_control[8] & 0x04)
|
||||
{
|
||||
videoram[offs ] = (videoram[offs ] & 0x3f) | (((data >> 0) & 3) << 6);
|
||||
videoram[offs+1] = (videoram[offs+1] & 0x3f) | (((data >> 2) & 3) << 6);
|
||||
videoram[offs+2] = (videoram[offs+2] & 0x3f) | (((data >> 4) & 3) << 6);
|
||||
videoram[offs+3] = (videoram[offs+3] & 0x3f) | (((data >> 6) & 3) << 6);
|
||||
}
|
||||
|
||||
/* A layer, green & blue bits */
|
||||
if (kangaroo_video_control[8] & 0x02)
|
||||
{
|
||||
videoram[offs ] = (videoram[offs ] & 0xfc) | (((data >> 0) & 3) << 0);
|
||||
videoram[offs+1] = (videoram[offs+1] & 0xfc) | (((data >> 2) & 3) << 0);
|
||||
videoram[offs+2] = (videoram[offs+2] & 0xfc) | (((data >> 4) & 3) << 0);
|
||||
videoram[offs+3] = (videoram[offs+3] & 0xfc) | (((data >> 6) & 3) << 0);
|
||||
}
|
||||
|
||||
/* A layer, Z & red bits */
|
||||
if (kangaroo_video_control[8] & 0x01)
|
||||
{
|
||||
videoram[offs ] = (videoram[offs ] & 0xf3) | (((data >> 0) & 3) << 2);
|
||||
videoram[offs+1] = (videoram[offs+1] & 0xf3) | (((data >> 2) & 3) << 2);
|
||||
videoram[offs+2] = (videoram[offs+2] & 0xf3) | (((data >> 4) & 3) << 2);
|
||||
videoram[offs+3] = (videoram[offs+3] & 0xf3) | (((data >> 6) & 3) << 2);
|
||||
}
|
||||
/* loop over height, then width */
|
||||
for (y = 0; y <= height; y++, dst += 256)
|
||||
for (x = 0; x <= width; x++)
|
||||
{
|
||||
UINT16 effdst = (dst + x) & 0x3fff;
|
||||
UINT16 effsrc = src++ & (gfxhalfsize - 1);
|
||||
videoram_write(effdst, gfxbase[0*gfxhalfsize + effsrc], mask & 0x05);
|
||||
videoram_write(effdst, gfxbase[1*gfxhalfsize + effsrc], mask & 0x0a);
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
@ -196,16 +166,18 @@ VIDEO_UPDATE( kangaroo )
|
||||
UINT8 effya = scrolly + (y ^ xora);
|
||||
UINT8 effxb = (x / 2) ^ xorb;
|
||||
UINT8 effyb = y ^ xorb;
|
||||
UINT8 pixa = videoram[effya * 256 + effxa] & 0x0f;
|
||||
UINT8 pixb = videoram[effyb * 256 + effxb] >> 4;
|
||||
UINT8 pixa = (videoram32[effya + 256 * (effxa / 4)] >> (8 * (effxa % 4) + 0)) & 0x0f;
|
||||
UINT8 pixb = (videoram32[effyb + 256 * (effxb / 4)] >> (8 * (effxb % 4) + 4)) & 0x0f;
|
||||
UINT8 finalpens;
|
||||
|
||||
/* for each layer, contribute bits if (a) enabled, and (b) either has priority or the opposite plane is 0 */
|
||||
finalpens = 0;
|
||||
if (enaa && (pria || pixb == 0))
|
||||
finalpens |= pixa;
|
||||
if (enab && (prib || pixa == 0))
|
||||
finalpens |= pixb;
|
||||
|
||||
/* store the first of two pixels, which is always full brightness */
|
||||
dest[x + 0] = pens[finalpens & 7];
|
||||
|
||||
/* KOS1 alternates at 5MHz, offset from the pixel clock by 1/2 clock */
|
||||
@ -222,6 +194,7 @@ VIDEO_UPDATE( kangaroo )
|
||||
finalpens |= pixb;
|
||||
}
|
||||
|
||||
/* store the second of two pixels, which is affected by KOS1 and the A/B masks */
|
||||
dest[x + 1] = pens[finalpens & 7];
|
||||
}
|
||||
}
|
||||
|
||||
Loading…
Reference in New Issue
Block a user