//
// Blitter core
//
-// by James L. Hammons
+// by James Hammons
+// (C) 2010 Underground Software
+//
+// JLH = James Hammons <jlhamm@acm.org>
+//
+// Who When What
+// --- ---------- -------------------------------------------------------------
+// JLH 01/16/2010 Created this log ;-)
+//
+
//
// I owe a debt of gratitude to Curt Vendel and to John Mathieson--to Curt
// for supplying the Oberon ASIC nets and to John for making them available
// to Curt. ;-) Without that excellent documentation which shows *exactly*
// what's going on inside the TOM chip, we'd all still be guessing as to how
// the wily blitter and other pieces of the Jaguar puzzle actually work.
+// Now how about those JERRY ASIC nets gentlemen...? [We have those now!] ;-)
//
-#include "jaguar.h"
#include "blitter.h"
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include "jaguar.h"
+#include "log.h"
+//#include "memory.h"
+
// Various conditional compilation goodies...
+//#define LOG_BLITS
+
//#define USE_ORIGINAL_BLITTER
//#define USE_MIDSUMMER_BLITTER
#define USE_MIDSUMMER_BLITTER_MKII
// Local global variables
int start_logging = 0;
-uint8 blitter_working = 0;
+uint8_t blitter_working = 0;
// Blitter register RAM (most of it is hidden from the user)
-static uint8 blitter_ram[0x100];
+static uint8_t blitter_ram[0x100];
// Other crapola
bool specialLog = false;
extern int effect_start;
extern int blit_start_log;
-void BlitterMidsummer(uint32 cmd);
+void BlitterMidsummer(uint32_t cmd);
void BlitterMidsummer2(void);
-#define REG(A) (((uint32)blitter_ram[(A)] << 24) | ((uint32)blitter_ram[(A)+1] << 16) \
- | ((uint32)blitter_ram[(A)+2] << 8) | (uint32)blitter_ram[(A)+3])
+#define REG(A) (((uint32_t)blitter_ram[(A)] << 24) | ((uint32_t)blitter_ram[(A)+1] << 16) \
+ | ((uint32_t)blitter_ram[(A)+2] << 8) | (uint32_t)blitter_ram[(A)+3])
#define WREG(A,D) (blitter_ram[(A)] = ((D)>>24)&0xFF, blitter_ram[(A)+1] = ((D)>>16)&0xFF, \
blitter_ram[(A)+2] = ((D)>>8)&0xFF, blitter_ram[(A)+3] = (D)&0xFF)
// Blitter registers (offsets from F02200)
-#define A1_BASE ((UINT32)0x00)
-#define A1_FLAGS ((UINT32)0x04)
-#define A1_CLIP ((UINT32)0x08) // Height and width values for clipping
-#define A1_PIXEL ((UINT32)0x0C) // Integer part of the pixel (Y.i and X.i)
-#define A1_STEP ((UINT32)0x10) // Integer part of the step
-#define A1_FSTEP ((UINT32)0x14) // Fractional part of the step
-#define A1_FPIXEL ((UINT32)0x18) // Fractional part of the pixel (Y.f and X.f)
-#define A1_INC ((UINT32)0x1C) // Integer part of the increment
-#define A1_FINC ((UINT32)0x20) // Fractional part of the increment
-#define A2_BASE ((UINT32)0x24)
-#define A2_FLAGS ((UINT32)0x28)
-#define A2_MASK ((UINT32)0x2C) // Modulo values for x and y (M.y and M.x)
-#define A2_PIXEL ((UINT32)0x30) // Integer part of the pixel (no fractional part for A2)
-#define A2_STEP ((UINT32)0x34) // Integer part of the step (no fractional part for A2)
-#define COMMAND ((UINT32)0x38)
-#define PIXLINECOUNTER ((UINT32)0x3C) // Inner & outer loop values
-#define SRCDATA ((UINT32)0x40)
-#define DSTDATA ((UINT32)0x48)
-#define DSTZ ((UINT32)0x50)
-#define SRCZINT ((UINT32)0x58)
-#define SRCZFRAC ((UINT32)0x60)
-#define PATTERNDATA ((UINT32)0x68)
-#define INTENSITYINC ((UINT32)0x70)
-#define ZINC ((UINT32)0x74)
-#define COLLISIONCTRL ((UINT32)0x78)
-#define PHRASEINT0 ((UINT32)0x7C)
-#define PHRASEINT1 ((UINT32)0x80)
-#define PHRASEINT2 ((UINT32)0x84)
-#define PHRASEINT3 ((UINT32)0x88)
-#define PHRASEZ0 ((UINT32)0x8C)
-#define PHRASEZ1 ((UINT32)0x90)
-#define PHRASEZ2 ((UINT32)0x94)
-#define PHRASEZ3 ((UINT32)0x98)
+#define A1_BASE ((uint32_t)0x00)
+#define A1_FLAGS ((uint32_t)0x04)
+#define A1_CLIP ((uint32_t)0x08) // Height and width values for clipping
+#define A1_PIXEL ((uint32_t)0x0C) // Integer part of the pixel (Y.i and X.i)
+#define A1_STEP ((uint32_t)0x10) // Integer part of the step
+#define A1_FSTEP ((uint32_t)0x14) // Fractional part of the step
+#define A1_FPIXEL ((uint32_t)0x18) // Fractional part of the pixel (Y.f and X.f)
+#define A1_INC ((uint32_t)0x1C) // Integer part of the increment
+#define A1_FINC ((uint32_t)0x20) // Fractional part of the increment
+#define A2_BASE ((uint32_t)0x24)
+#define A2_FLAGS ((uint32_t)0x28)
+#define A2_MASK ((uint32_t)0x2C) // Modulo values for x and y (M.y and M.x)
+#define A2_PIXEL ((uint32_t)0x30) // Integer part of the pixel (no fractional part for A2)
+#define A2_STEP ((uint32_t)0x34) // Integer part of the step (no fractional part for A2)
+#define COMMAND ((uint32_t)0x38)
+#define PIXLINECOUNTER ((uint32_t)0x3C) // Inner & outer loop values
+#define SRCDATA ((uint32_t)0x40)
+#define DSTDATA ((uint32_t)0x48)
+#define DSTZ ((uint32_t)0x50)
+#define SRCZINT ((uint32_t)0x58)
+#define SRCZFRAC ((uint32_t)0x60)
+#define PATTERNDATA ((uint32_t)0x68)
+#define INTENSITYINC ((uint32_t)0x70)
+#define ZINC ((uint32_t)0x74)
+#define COLLISIONCTRL ((uint32_t)0x78)
+#define PHRASEINT0 ((uint32_t)0x7C)
+#define PHRASEINT1 ((uint32_t)0x80)
+#define PHRASEINT2 ((uint32_t)0x84)
+#define PHRASEINT3 ((uint32_t)0x88)
+#define PHRASEZ0 ((uint32_t)0x8C)
+#define PHRASEZ1 ((uint32_t)0x90)
+#define PHRASEZ2 ((uint32_t)0x94)
+#define PHRASEZ3 ((uint32_t)0x98)
// Blitter command bits
//Put 'em back, once we fix the problem!!! [KO]
// 1 bpp pixel read
#define PIXEL_SHIFT_1(a) (((~a##_x) >> 16) & 7)
-#define PIXEL_OFFSET_1(a) (((((UINT32)a##_y >> 16) * a##_width / 8) + (((UINT32)a##_x >> 19) & ~7)) * (1 + a##_pitch) + (((UINT32)a##_x >> 19) & 7))
+#define PIXEL_OFFSET_1(a) (((((uint32_t)a##_y >> 16) * a##_width / 8) + (((uint32_t)a##_x >> 19) & ~7)) * (1 + a##_pitch) + (((uint32_t)a##_x >> 19) & 7))
#define READ_PIXEL_1(a) ((JaguarReadByte(a##_addr+PIXEL_OFFSET_1(a), BLITTER) >> PIXEL_SHIFT_1(a)) & 0x01)
//#define READ_PIXEL_1(a) ((JaguarReadByte(a##_addr+PIXEL_OFFSET_1(a)) >> PIXEL_SHIFT_1(a)) & 0x01)
// 2 bpp pixel read
#define PIXEL_SHIFT_2(a) (((~a##_x) >> 15) & 6)
-#define PIXEL_OFFSET_2(a) (((((UINT32)a##_y >> 16) * a##_width / 4) + (((UINT32)a##_x >> 18) & ~7)) * (1 + a##_pitch) + (((UINT32)a##_x >> 18) & 7))
+#define PIXEL_OFFSET_2(a) (((((uint32_t)a##_y >> 16) * a##_width / 4) + (((uint32_t)a##_x >> 18) & ~7)) * (1 + a##_pitch) + (((uint32_t)a##_x >> 18) & 7))
#define READ_PIXEL_2(a) ((JaguarReadByte(a##_addr+PIXEL_OFFSET_2(a), BLITTER) >> PIXEL_SHIFT_2(a)) & 0x03)
//#define READ_PIXEL_2(a) ((JaguarReadByte(a##_addr+PIXEL_OFFSET_2(a)) >> PIXEL_SHIFT_2(a)) & 0x03)
// 4 bpp pixel read
#define PIXEL_SHIFT_4(a) (((~a##_x) >> 14) & 4)
-#define PIXEL_OFFSET_4(a) (((((UINT32)a##_y >> 16) * (a##_width/2)) + (((UINT32)a##_x >> 17) & ~7)) * (1 + a##_pitch) + (((UINT32)a##_x >> 17) & 7))
+#define PIXEL_OFFSET_4(a) (((((uint32_t)a##_y >> 16) * (a##_width/2)) + (((uint32_t)a##_x >> 17) & ~7)) * (1 + a##_pitch) + (((uint32_t)a##_x >> 17) & 7))
#define READ_PIXEL_4(a) ((JaguarReadByte(a##_addr+PIXEL_OFFSET_4(a), BLITTER) >> PIXEL_SHIFT_4(a)) & 0x0f)
//#define READ_PIXEL_4(a) ((JaguarReadByte(a##_addr+PIXEL_OFFSET_4(a)) >> PIXEL_SHIFT_4(a)) & 0x0f)
// 8 bpp pixel read
-#define PIXEL_OFFSET_8(a) (((((UINT32)a##_y >> 16) * a##_width) + (((UINT32)a##_x >> 16) & ~7)) * (1 + a##_pitch) + (((UINT32)a##_x >> 16) & 7))
+#define PIXEL_OFFSET_8(a) (((((uint32_t)a##_y >> 16) * a##_width) + (((uint32_t)a##_x >> 16) & ~7)) * (1 + a##_pitch) + (((uint32_t)a##_x >> 16) & 7))
#define READ_PIXEL_8(a) (JaguarReadByte(a##_addr+PIXEL_OFFSET_8(a), BLITTER))
//#define READ_PIXEL_8(a) (JaguarReadByte(a##_addr+PIXEL_OFFSET_8(a)))
// 16 bpp pixel read
-#define PIXEL_OFFSET_16(a) (((((UINT32)a##_y >> 16) * a##_width) + (((UINT32)a##_x >> 16) & ~3)) * (1 + a##_pitch) + (((UINT32)a##_x >> 16) & 3))
+#define PIXEL_OFFSET_16(a) (((((uint32_t)a##_y >> 16) * a##_width) + (((uint32_t)a##_x >> 16) & ~3)) * (1 + a##_pitch) + (((uint32_t)a##_x >> 16) & 3))
#define READ_PIXEL_16(a) (JaguarReadWord(a##_addr+(PIXEL_OFFSET_16(a)<<1), BLITTER))
//#define READ_PIXEL_16(a) (JaguarReadWord(a##_addr+(PIXEL_OFFSET_16(a)<<1)))
// 32 bpp pixel read
-#define PIXEL_OFFSET_32(a) (((((UINT32)a##_y >> 16) * a##_width) + (((UINT32)a##_x >> 16) & ~1)) * (1 + a##_pitch) + (((UINT32)a##_x >> 16) & 1))
+#define PIXEL_OFFSET_32(a) (((((uint32_t)a##_y >> 16) * a##_width) + (((uint32_t)a##_x >> 16) & ~1)) * (1 + a##_pitch) + (((uint32_t)a##_x >> 16) & 1))
#define READ_PIXEL_32(a) (JaguarReadLong(a##_addr+(PIXEL_OFFSET_32(a)<<2), BLITTER))
//#define READ_PIXEL_32(a) (JaguarReadLong(a##_addr+(PIXEL_OFFSET_32(a)<<2)))
//#define WRITE_ZDATA_16(a,d) { JaguarWriteWord(a##_addr+(ZDATA_OFFSET_16(a)<<1), d); }
// z data write
-#define WRITE_ZDATA(a,f,d) WRITE_ZDATA_16(a,d);
+#define WRITE_ZDATA(a,f,d) WRITE_ZDATA_16(a,d);
// 1 bpp r data read
-#define READ_RDATA_1(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 19) & 0x04))) >> (((UINT32)a##_x >> 16) & 0x1F)) & 0x0001 : (REG(r) & 0x0001))
+#define READ_RDATA_1(r,a,p) ((p) ? ((REG(r+(((uint32_t)a##_x >> 19) & 0x04))) >> (((uint32_t)a##_x >> 16) & 0x1F)) & 0x0001 : (REG(r) & 0x0001))
// 2 bpp r data read
-#define READ_RDATA_2(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 18) & 0x04))) >> (((UINT32)a##_x >> 15) & 0x3E)) & 0x0003 : (REG(r) & 0x0003))
+#define READ_RDATA_2(r,a,p) ((p) ? ((REG(r+(((uint32_t)a##_x >> 18) & 0x04))) >> (((uint32_t)a##_x >> 15) & 0x3E)) & 0x0003 : (REG(r) & 0x0003))
// 4 bpp r data read
-#define READ_RDATA_4(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 17) & 0x04))) >> (((UINT32)a##_x >> 14) & 0x28)) & 0x000F : (REG(r) & 0x000F))
+#define READ_RDATA_4(r,a,p) ((p) ? ((REG(r+(((uint32_t)a##_x >> 17) & 0x04))) >> (((uint32_t)a##_x >> 14) & 0x28)) & 0x000F : (REG(r) & 0x000F))
// 8 bpp r data read
-#define READ_RDATA_8(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 16) & 0x04))) >> (((UINT32)a##_x >> 13) & 0x18)) & 0x00FF : (REG(r) & 0x00FF))
+#define READ_RDATA_8(r,a,p) ((p) ? ((REG(r+(((uint32_t)a##_x >> 16) & 0x04))) >> (((uint32_t)a##_x >> 13) & 0x18)) & 0x00FF : (REG(r) & 0x00FF))
// 16 bpp r data read
-#define READ_RDATA_16(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 15) & 0x04))) >> (((UINT32)a##_x >> 12) & 0x10)) & 0xFFFF : (REG(r) & 0xFFFF))
+#define READ_RDATA_16(r,a,p) ((p) ? ((REG(r+(((uint32_t)a##_x >> 15) & 0x04))) >> (((uint32_t)a##_x >> 12) & 0x10)) & 0xFFFF : (REG(r) & 0xFFFF))
// 32 bpp r data read
-#define READ_RDATA_32(r,a,p) ((p) ? REG(r+(((UINT32)a##_x >> 14) & 0x04)) : REG(r))
+#define READ_RDATA_32(r,a,p) ((p) ? REG(r+(((uint32_t)a##_x >> 14) & 0x04)) : REG(r))
// register data read
#define READ_RDATA(r,a,f,p) (\
//#define WRITE_PIXEL_16(a,d) { JaguarWriteWord(a##_addr+(PIXEL_OFFSET_16(a)<<1), d); if (specialLog) WriteLog("Pixel write address: %08X\n", a##_addr+(PIXEL_OFFSET_16(a)<<1)); }
// 32 bpp pixel write
-#define WRITE_PIXEL_32(a,d) { JaguarWriteLong(a##_addr+(PIXEL_OFFSET_32(a)<<2), d, BLITTER); }
-//#define WRITE_PIXEL_32(a,d) { JaguarWriteLong(a##_addr+(PIXEL_OFFSET_32(a)<<2), d); }
+#define WRITE_PIXEL_32(a,d) { JaguarWriteLong(a##_addr+(PIXEL_OFFSET_32(a)<<2), d, BLITTER); }
+//#define WRITE_PIXEL_32(a,d) { JaguarWriteLong(a##_addr+(PIXEL_OFFSET_32(a)<<2), d); }
// pixel write
#define WRITE_PIXEL(a,f,d) {\
// as a floating point bit pattern being followed by a number of zeroes. So, e.g., 001101 translates to
// 1.01 (the "1." being implied) x (2 ^ 3) or 1010 -> 10 in base 10 (i.e., 1.01 with the decimal place
// being shifted to the right 3 places).
-/*static uint32 blitter_scanline_width[48] =
-{
+/*static uint32_t blitter_scanline_width[48] =
+{
0, 0, 0, 0, // Note: This would really translate to 1, 1, 1, 1
2, 0, 0, 0,
4, 0, 6, 0,
2048, 2560, 3072, 3584
};//*/
-//static uint8 * tom_ram_8;
-//static uint8 * paletteRam;
-static uint8 src;
-static uint8 dst;
-static uint8 misc;
-static uint8 a1ctl;
-static uint8 mode;
-static uint8 ity;
-static uint8 zop;
-static uint8 op;
-static uint8 ctrl;
-static uint32 a1_addr;
-static uint32 a2_addr;
-static int32 a1_zoffs;
-static int32 a2_zoffs;
-static uint32 xadd_a1_control;
-static uint32 xadd_a2_control;
-static int32 a1_pitch;
-static int32 a2_pitch;
-static uint32 n_pixels;
-static uint32 n_lines;
-static int32 a1_x;
-static int32 a1_y;
-static int32 a1_width;
-static int32 a2_x;
-static int32 a2_y;
-static int32 a2_width;
-static int32 a2_mask_x;
-static int32 a2_mask_y;
-static int32 a1_xadd;
-static int32 a1_yadd;
-static int32 a2_xadd;
-static int32 a2_yadd;
-static uint8 a1_phrase_mode;
-static uint8 a2_phrase_mode;
-static int32 a1_step_x = 0;
-static int32 a1_step_y = 0;
-static int32 a2_step_x = 0;
-static int32 a2_step_y = 0;
-static uint32 outer_loop;
-static uint32 inner_loop;
-static uint32 a2_psize;
-static uint32 a1_psize;
-static uint32 gouraud_add;
-//static uint32 gouraud_data;
-//static uint16 gint[4];
-//static uint16 gfrac[4];
-//static uint8 gcolour[4];
+//static uint8_t * tom_ram_8;
+//static uint8_t * paletteRam;
+static uint8_t src;
+static uint8_t dst;
+static uint8_t misc;
+static uint8_t a1ctl;
+static uint8_t mode;
+static uint8_t ity;
+static uint8_t zop;
+static uint8_t op;
+static uint8_t ctrl;
+static uint32_t a1_addr;
+static uint32_t a2_addr;
+static int32_t a1_zoffs;
+static int32_t a2_zoffs;
+static uint32_t xadd_a1_control;
+static uint32_t xadd_a2_control;
+static int32_t a1_pitch;
+static int32_t a2_pitch;
+static uint32_t n_pixels;
+static uint32_t n_lines;
+static int32_t a1_x;
+static int32_t a1_y;
+static int32_t a1_width;
+static int32_t a2_x;
+static int32_t a2_y;
+static int32_t a2_width;
+static int32_t a2_mask_x;
+static int32_t a2_mask_y;
+static int32_t a1_xadd;
+static int32_t a1_yadd;
+static int32_t a2_xadd;
+static int32_t a2_yadd;
+static uint8_t a1_phrase_mode;
+static uint8_t a2_phrase_mode;
+static int32_t a1_step_x = 0;
+static int32_t a1_step_y = 0;
+static int32_t a2_step_x = 0;
+static int32_t a2_step_y = 0;
+static uint32_t outer_loop;
+static uint32_t inner_loop;
+static uint32_t a2_psize;
+static uint32_t a1_psize;
+static uint32_t gouraud_add;
+//static uint32_t gouraud_data;
+//static uint16_t gint[4];
+//static uint16_t gfrac[4];
+//static uint8_t gcolour[4];
static int gd_i[4];
static int gd_c[4];
static int gd_ia, gd_ca;
static int colour_index = 0;
-static int32 zadd;
-static uint32 z_i[4];
+static int32_t zadd;
+static uint32_t z_i[4];
-static int32 a1_clip_x, a1_clip_y;
+static int32_t a1_clip_x, a1_clip_y;
// In the spirit of "get it right first, *then* optimize" I've taken the liberty
// of removing all the unnecessary code caching. If it turns out to be a good way
//
// Generic blit handler
//
-void blitter_generic(uint32 cmd)
+void blitter_generic(uint32_t cmd)
{
/*
Blit! (0018FA70 <- 008DDC40) count: 2 x 13, A1/2_FLAGS: 00014218/00013C18 [cmd: 1401060C]
- CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
+ CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
A1 step values: -2 (X), 1 (Y)
A2 step values: -1 (X), 1 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 320 (21), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
/*if (cmd == 0x1401060C && blit_start_log)
specialLog = true;//*/
//Testing only!
-//uint32 logGo = ((cmd == 0x01800E01 && REG(A1_BASE) == 0x898000) ? 1 : 0);
- uint32 srcdata, srczdata, dstdata, dstzdata, writedata, inhibit;
- uint32 bppSrc = (DSTA2 ? 1 << ((REG(A1_FLAGS) >> 3) & 0x07) : 1 << ((REG(A2_FLAGS) >> 3) & 0x07));
+//uint32_t logGo = ((cmd == 0x01800E01 && REG(A1_BASE) == 0x898000) ? 1 : 0);
+ uint32_t srcdata, srczdata, dstdata, dstzdata, writedata, inhibit;
+ uint32_t bppSrc = (DSTA2 ? 1 << ((REG(A1_FLAGS) >> 3) & 0x07) : 1 << ((REG(A2_FLAGS) >> 3) & 0x07));
if (specialLog)
{
{
WriteLog(" A1_X/Y = %08X/%08X, A2_X/Y = %08X/%08X\n", a1_x, a1_y, a2_x, a2_y);
}
- uint32 a1_start = a1_x, a2_start = a2_x, bitPos = 0;
+ uint32_t a1_start = a1_x, a2_start = a2_x, bitPos = 0;
//Kludge for Hover Strike...
//I wonder if this kludge is in conjunction with the SRCENX down below...
srczdata = READ_RDATA(SRCZINT, a2, REG(A2_FLAGS), a2_phrase_mode);
}
- // load dst data and Z
+ // load dst data and Z
if (DSTEN)
{
dstdata = READ_PIXEL(a1, REG(A1_FLAGS));
inhibit = 1;
}//*/
- if (GOURZ)
+ if (GOURZ)
srczdata = z_i[colour_index] >> 16;
// apply z comparator
if (Z_OP_INF && srczdata < dstzdata) inhibit = 1;
if (Z_OP_EQU && srczdata == dstzdata) inhibit = 1;
if (Z_OP_SUP && srczdata > dstzdata) inhibit = 1;
-
+
// apply data comparator
// Note: DCOMPEN only works in 8/16 bpp modes! !!! FIX !!!
// Does BCOMPEN only work in 1 bpp mode???
//and *then* we can do the bit stepping from there the way it's *supposed* to be done... !!! FIX !!!
//[DONE]
{
- uint32 pixShift = (~bitPos) & (bppSrc - 1);
+ uint32_t pixShift = (~bitPos) & (bppSrc - 1);
srcdata = (srcdata >> pixShift) & 0x01;
bitPos++;
Interesting (Hover Strike--large letter):
Blit! (0018FA70 <- 008DDC40) count: 2 x 13, A1/2_FLAGS: 00014218/00013C18 [cmd: 1401060C]
- CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
+ CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
A1 step values: -2 (X), 1 (Y)
A2 step values: -1 (X), 1 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 320 (21), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A1 x/y: 100/12, A2 x/y: 106/0 Pattern: 000000F300000000
Blit! (0018FA70 <- 008DDC40) count: 8 x 13, A1/2_FLAGS: 00014218/00013C18 [cmd: 1401060C]
- CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
+ CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
A1 step values: -8 (X), 1 (Y)
A2 step values: -1 (X), 1 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 320 (21), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A1 x/y: 102/12, A2 x/y: 107/0 Pattern: 000000F300000000
Blit! (0018FA70 <- 008DDC40) count: 1 x 13, A1/2_FLAGS: 00014218/00013C18 [cmd: 1401060C]
- CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
+ CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
A1 step values: -1 (X), 1 (Y)
A2 step values: -1 (X), 1 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 320 (21), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A1 x/y: 118/12, A2 x/y: 70/0 Pattern: 000000F300000000
Blit! (0018FA70 <- 008DDC40) count: 8 x 13, A1/2_FLAGS: 00014218/00013C18 [cmd: 1401060C]
- CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
+ CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
A1 step values: -8 (X), 1 (Y)
A2 step values: -1 (X), 1 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 320 (21), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A1 x/y: 119/12, A2 x/y: 71/0 Pattern: 000000F300000000
Blit! (0018FA70 <- 008DDC40) count: 1 x 13, A1/2_FLAGS: 00014218/00013C18 [cmd: 1401060C]
- CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
+ CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
A1 step values: -1 (X), 1 (Y)
A2 step values: -1 (X), 1 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 320 (21), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A1 x/y: 127/12, A2 x/y: 66/0 Pattern: 000000F300000000
Blit! (0018FA70 <- 008DDC40) count: 8 x 13, A1/2_FLAGS: 00014218/00013C18 [cmd: 1401060C]
- CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
+ CMD -> src: SRCENX dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl: BCOMPEN BKGWREN
A1 step values: -8 (X), 1 (Y)
A2 step values: -1 (X), 1 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 320 (21), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
// compare source pixel with pattern pixel
/*
Blit! (000B8250 <- 0012C3A0) count: 16 x 1, A1/2_FLAGS: 00014420/00012000 [cmd: 05810001]
- CMD -> src: SRCEN dst: misc: a1ctl: mode: ity: PATDSEL z-op: op: LFU_REPLACE ctrl: BCOMPEN
+ CMD -> src: SRCEN dst: misc: a1ctl: mode: ity: PATDSEL z-op: op: LFU_REPLACE ctrl: BCOMPEN
A1 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 384 (22), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A2 -> pitch: 1 phrases, depth: 1bpp, z-off: 0, width: 16 (10), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
x/y: 0/20
/* if (srcdata == READ_RDATA(PATTERNDATA, a2, REG(A2_FLAGS), a2_phrase_mode))
// if (srcdata != READ_RDATA(PATTERNDATA, a2, REG(A2_FLAGS), a2_phrase_mode))
inhibit = 1;//*/
-/* uint32 A2bpp = 1 << ((REG(A2_FLAGS) >> 3) & 0x07);
+/* uint32_t A2bpp = 1 << ((REG(A2_FLAGS) >> 3) & 0x07);
if (A2bpp == 1 || A2bpp == 16 || A2bpp == 8)
inhibit = (srcdata == 0 ? 1: 0);
// inhibit = !srcdata;
// compute the write data and store
if (!inhibit)
- {
+ {
// Houston, we have a problem...
// Look here, at PATDSEL and GOURD. If both are active (as they are on the BIOS intro), then there's
// a conflict! E.g.:
//Blit! (00100000 <- 000095D0) count: 3 x 1, A1/2_FLAGS: 00014220/00004020 [cmd: 00011008]
-// CMD -> src: dst: DSTEN misc: a1ctl: mode: GOURD ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
+// CMD -> src: dst: DSTEN misc: a1ctl: mode: GOURD ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
// A1 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 320 (21), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
// A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 256 (20), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
// A1 x/y: 90/171, A2 x/y: 808/0 Pattern: 776D770077007700
Hover Strike ADDDSEL blit:
Blit! (00098D90 <- 0081DDC0) count: 320 x 287, A1/2_FLAGS: 00004220/00004020 [cmd: 00020208]
- CMD -> src: dst: DSTEN misc: a1ctl: UPDA1 mode: ity: ADDDSEL z-op: op: LFU_CLEAR ctrl:
+ CMD -> src: dst: DSTEN misc: a1ctl: UPDA1 mode: ity: ADDDSEL z-op: op: LFU_CLEAR ctrl:
A1 step values: -320 (X), 1 (Y)
A1 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 256 (20), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
if (!TOPBEN)
{
//This is correct now, but slow...
- int16 s = (srcdata & 0xFF) | (srcdata & 0x80 ? 0xFF00 : 0x0000),
+ int16_t s = (srcdata & 0xFF) | (srcdata & 0x80 ? 0xFF00 : 0x0000),
d = dstdata & 0xFF;
- int16 sum = s + d;
+ int16_t sum = s + d;
if (sum < 0)
writedata = 0x00;
else if (sum > 0xFF)
writedata = 0xFF;
else
- writedata = (uint32)sum;
+ writedata = (uint32_t)sum;
}
//This doesn't seem right... Looks like it would muck up the low byte... !!! FIX !!!
//According to JTRM, this is part of the four things the blitter does with the write data (the other
//three being PATDSEL, ADDDSEL, and LFU (default). I'm not sure which gets precedence, this or PATDSEL
//(see above blit example)...
- if (GOURD)
+ if (GOURD)
writedata = ((gd_c[colour_index]) << 8) | (gd_i[colour_index] >> 16);
- if (SRCSHADE)
+ if (SRCSHADE)
{
int intensity = srcdata & 0xFF;
int ia = gd_ia >> 16;
{
/*if (((REG(A1_FLAGS) >> 3) & 0x07) == 5)
{
- uint32 offset = a1_addr+(PIXEL_OFFSET_32(a1)<<2);
-// (((((UINT32)a##_y >> 16) * a##_width) + (((UINT32)a##_x >> 16) & ~1)) * (1 + a##_pitch) + (((UINT32)a##_x >> 16) & 1))
+ uint32_t offset = a1_addr+(PIXEL_OFFSET_32(a1)<<2);
+// (((((uint32_t)a##_y >> 16) * a##_width) + (((uint32_t)a##_x >> 16) & ~1)) * (1 + a##_pitch) + (((uint32_t)a##_x >> 16) & 1))
if ((offset >= 0x1FF020 && offset <= 0x1FF03F) || (offset >= 0x1FF820 && offset <= 0x1FF83F))
WriteLog("32bpp pixel write: A1 Phrase mode --> ");
}//*/
srczdata = READ_RDATA(SRCZINT, a1, REG(A1_FLAGS), a1_phrase_mode);
}
- // load dst data and Z
+ // load dst data and Z
if (DSTEN)
{
dstdata = READ_PIXEL(a2, REG(A2_FLAGS));
dstzdata = READ_RDATA(DSTZ, a2, REG(A2_FLAGS), a2_phrase_mode);
}
- if (GOURZ)
+ if (GOURZ)
srczdata = z_i[colour_index] >> 16;
// apply z comparator
if (Z_OP_INF && srczdata < dstzdata) inhibit = 1;
if (Z_OP_EQU && srczdata == dstzdata) inhibit = 1;
if (Z_OP_SUP && srczdata > dstzdata) inhibit = 1;
-
+
// apply data comparator
//NOTE: The bit comparator (BCOMPEN) is NOT the same at the data comparator!
if (DCOMPEN | BCOMPEN)
// AvP: This is causing blocks to be written instead of bit patterns...
// Works now...
// NOTE: We really should separate out the BCOMPEN & DCOMPEN stuff!
-/* uint32 A1bpp = 1 << ((REG(A1_FLAGS) >> 3) & 0x07);
+/* uint32_t A1bpp = 1 << ((REG(A1_FLAGS) >> 3) & 0x07);
if (A1bpp == 1 || A1bpp == 16 || A1bpp == 8)
inhibit = (srcdata == 0 ? 1: 0);
else
// if (a1_phrase_mode || a2_phrase_mode)
// inhibit = !inhibit;
}
-
+
if (CLIPA1)
{
inhibit |= (((a1_x >> 16) < a1_clip_x && (a1_x >> 16) >= 0
// compute the write data and store
if (!inhibit)
- {
+ {
if (PATDSEL)
{
// use pattern data for write data
writedata |= srcdata & dstdata;
}
- if (GOURD)
+ if (GOURD)
writedata = ((gd_c[colour_index]) << 8) | (gd_i[colour_index] >> 16);
- if (SRCSHADE)
+ if (SRCSHADE)
{
int intensity = srcdata & 0xFF;
int ia = gd_ia >> 16;
{
/*if (logGo)
{
- uint32 offset = a2_addr+(PIXEL_OFFSET_16(a2)<<1);
-// (((((UINT32)a##_y >> 16) * a##_width) + (((UINT32)a##_x >> 16) & ~1)) * (1 + a##_pitch) + (((UINT32)a##_x >> 16) & 1))
+ uint32_t offset = a2_addr+(PIXEL_OFFSET_16(a2)<<1);
+// (((((uint32_t)a##_y >> 16) * a##_width) + (((uint32_t)a##_x >> 16) & ~1)) * (1 + a##_pitch) + (((uint32_t)a##_x >> 16) & 1))
WriteLog("[%08X:%04X] ", offset, writedata);
}//*/
// write to the destination
gd_i[colour_index] += gd_ia;
//Hmm, this doesn't seem to do anything...
//But it is correct according to the JTRM...!
-if ((int32)gd_i[colour_index] < 0)
+if ((int32_t)gd_i[colour_index] < 0)
gd_i[colour_index] = 0;
if (gd_i[colour_index] > 0x00FFFFFF)
gd_i[colour_index] = 0x00FFFFFF;//*/
gd_c[colour_index] += gd_ca;
-if ((int32)gd_c[colour_index] < 0)
+if ((int32_t)gd_c[colour_index] < 0)
gd_c[colour_index] = 0;
if (gd_c[colour_index] > 0x000000FF)
gd_c[colour_index] = 0x000000FF;//*/
; O
Blit! (00110000 <- 0010B2A8) count: 12 x 12, A1/2_FLAGS: 000042E2/00000020 [cmd: 09800609]
- CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
+ CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
A1 step values: -15 (X), 1 (Y)
A2 step values: -4 (X), 0 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
*/
//NOTE: The way to fix the CD BIOS is to uncomment below and comment the stuff after
// the phrase mode mucking around. But it fucks up everything else...
-#define SCREWY_CD_DEPENDENT
+//#define SCREWY_CD_DEPENDENT
#ifdef SCREWY_CD_DEPENDENT
a1_x += a1_step_x;
a1_y += a1_step_y;
{
// Bump the pointer to the next phrase boundary
// Even though it works, this is crappy... Clean it up!
- uint32 size = 64 / a1_psize;
+ uint32_t size = 64 / a1_psize;
// Crappy kludge... ('aligning' source to destination)
if (a2_phrase_mode && DSTA2)
{
- uint32 extra = (a2_start >> 16) % size;
+ uint32_t extra = (a2_start >> 16) % size;
a1_x += extra << 16;
}
- uint32 newx = (a1_x >> 16) / size;
- uint32 newxrem = (a1_x >> 16) % size;
+ uint32_t newx = (a1_x >> 16) / size;
+ uint32_t newxrem = (a1_x >> 16) % size;
a1_x &= 0x0000FFFF;
a1_x |= (((newx + (newxrem == 0 ? 0 : 1)) * size) & 0xFFFF) << 16;
}//*/
{
// Bump the pointer to the next phrase boundary
// Even though it works, this is crappy... Clean it up!
- uint32 size = 64 / a1_psize;
+ uint32_t size = 64 / a1_psize;
// Crappy kludge... ('aligning' source to destination)
if (a2_phrase_mode && DSTA2)
{
- uint32 extra = (a2_start >> 16) % size;
+ uint32_t extra = (a2_start >> 16) % size;
a1_x += extra << 16;
}
- uint32 pixelSize = (size - 1) << 16;
+ uint32_t pixelSize = (size - 1) << 16;
a1_x = (a1_x + pixelSize) & ~pixelSize;
}
{
// Bump the pointer to the next phrase boundary
// Even though it works, this is crappy... Clean it up!
- uint32 size = 64 / a2_psize;
+ uint32_t size = 64 / a2_psize;
// Crappy kludge... ('aligning' source to destination)
// Prolly should do this for A1 channel as well... [DONE]
if (a1_phrase_mode && !DSTA2)
{
- uint32 extra = (a1_start >> 16) % size;
+ uint32_t extra = (a1_start >> 16) % size;
a2_x += extra << 16;
}
- uint32 newx = (a2_x >> 16) / size;
- uint32 newxrem = (a2_x >> 16) % size;
+ uint32_t newx = (a2_x >> 16) / size;
+ uint32_t newxrem = (a2_x >> 16) % size;
a2_x &= 0x0000FFFF;
a2_x |= (((newx + (newxrem == 0 ? 0 : 1)) * size) & 0xFFFF) << 16;
}//*/
{
// Bump the pointer to the next phrase boundary
// Even though it works, this is crappy... Clean it up!
- uint32 size = 64 / a2_psize;
+ uint32_t size = 64 / a2_psize;
// Crappy kludge... ('aligning' source to destination)
// Prolly should do this for A1 channel as well... [DONE]
if (a1_phrase_mode && !DSTA2)
{
- uint32 extra = (a1_start >> 16) % size;
+ uint32_t extra = (a1_start >> 16) % size;
a2_x += extra << 16;
}
- uint32 pixelSize = (size - 1) << 16;
+ uint32_t pixelSize = (size - 1) << 16;
a2_x = (a2_x + pixelSize) & ~pixelSize;
}
a2_y += a2_step_y;//*/
#endif
}
-
- // write values back to registers
+
+ // write values back to registers
WREG(A1_PIXEL, (a1_y & 0xFFFF0000) | ((a1_x >> 16) & 0xFFFF));
WREG(A1_FPIXEL, (a1_y << 16) | (a1_x & 0xFFFF));
WREG(A2_PIXEL, (a2_y & 0xFFFF0000) | ((a2_x >> 16) & 0xFFFF));
specialLog = false;
}
-void blitter_blit(uint32 cmd)
+void blitter_blit(uint32_t cmd)
{
//Apparently this is doing *something*, just not sure exactly what...
/*if (cmd == 0x41802E01)
WriteLog("Is: %08X\n", cmd);
}//*/
- uint32 pitchValue[4] = { 0, 1, 3, 2 };
+ uint32_t pitchValue[4] = { 0, 1, 3, 2 };
colour_index = 0;
src = cmd & 0x07;
dst = (cmd >> 3) & 0x07;
a1_zoffs = (REG(A1_FLAGS) >> 6) & 7;
a2_zoffs = (REG(A2_FLAGS) >> 6) & 7;
-
+
xadd_a1_control = (REG(A1_FLAGS) >> 16) & 0x03;
xadd_a2_control = (REG(A2_FLAGS) >> 16) & 0x03;
// a1_width = blitter_scanline_width[((REG(A1_FLAGS) & 0x00007E00) >> 9)];
// According to JTRM, this must give a *whole number* of phrases in the current
// pixel size (this means the lookup above is WRONG)... !!! FIX !!!
- UINT32 m = (REG(A1_FLAGS) >> 9) & 0x03, e = (REG(A1_FLAGS) >> 11) & 0x0F;
+ uint32_t m = (REG(A1_FLAGS) >> 9) & 0x03, e = (REG(A1_FLAGS) >> 11) & 0x0F;
a1_width = ((0x04 | m) << e) >> 2;//*/
a2_x = (REG(A2_PIXEL) & 0x0000FFFF) << 16;
// add pixelsize (1) to X
a1_xadd = 1 << 16;
break;
- case XADD0:
+ case XADD0:
// add zero (for those nice vertical lines)
a1_xadd = 0;
break;
// A1 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 128 (1C), addctl: XADDINC YADD1 XSIGNADD YSIGNADD
// A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 320 (21), addctl: XADD0 YADD1 XSIGNADD YSIGNADD
//if (YADD1_A1 && YADD1_A2 && xadd_a2_control == XADD0 && xadd_a1_control == XADDINC)// &&
-// UINT32 a1f = REG(A1_FLAGS), a2f = REG(A2_FLAGS);
+// uint32_t a1f = REG(A1_FLAGS), a2f = REG(A2_FLAGS);
//Ok, so this ISN'T it... Prolly the XADDPHR code above that's doing it...
//if (REG(A1_FLAGS) == 0x00073820 && REG(A2_FLAGS) == 0x00064220 && cmd == 0x41802801)
// A1 x/y: 14368/7, A2 x/y: 150/36
// add pixelsize (1) to X
a2_xadd = 1 << 16;
break;
- case XADD0:
+ case XADD0:
// add zero (for those nice vertical lines)
a2_xadd = 0;
break;
if (GOURD || GOURZ || SRCSHADE)
{
gd_c[0] = blitter_ram[PATTERNDATA + 6];
- gd_i[0] = ((uint32)blitter_ram[PATTERNDATA + 7] << 16)
- | ((uint32)blitter_ram[SRCDATA + 6] << 8) | blitter_ram[SRCDATA + 1];
+ gd_i[0] = ((uint32_t)blitter_ram[PATTERNDATA + 7] << 16)
+ | ((uint32_t)blitter_ram[SRCDATA + 6] << 8) | blitter_ram[SRCDATA + 7];
gd_c[1] = blitter_ram[PATTERNDATA + 4];
- gd_i[1] = ((uint32)blitter_ram[PATTERNDATA + 5] << 16)
- | ((uint32)blitter_ram[SRCDATA + 4] << 8) | blitter_ram[SRCDATA + 3];
+ gd_i[1] = ((uint32_t)blitter_ram[PATTERNDATA + 5] << 16)
+ | ((uint32_t)blitter_ram[SRCDATA + 4] << 8) | blitter_ram[SRCDATA + 5];
gd_c[2] = blitter_ram[PATTERNDATA + 2];
- gd_i[2] = ((uint32)blitter_ram[PATTERNDATA + 3] << 16)
- | ((uint32)blitter_ram[SRCDATA + 2] << 8) | blitter_ram[SRCDATA + 5];
+ gd_i[2] = ((uint32_t)blitter_ram[PATTERNDATA + 3] << 16)
+ | ((uint32_t)blitter_ram[SRCDATA + 2] << 8) | blitter_ram[SRCDATA + 3];
gd_c[3] = blitter_ram[PATTERNDATA + 0];
- gd_i[3] = ((uint32)blitter_ram[PATTERNDATA + 1] << 16)
- | ((uint32)blitter_ram[SRCDATA + 0] << 8) | blitter_ram[SRCDATA + 7];
+ gd_i[3] = ((uint32_t)blitter_ram[PATTERNDATA + 1] << 16)
+ | ((uint32_t)blitter_ram[SRCDATA + 0] << 8) | blitter_ram[SRCDATA + 1];
gouraud_add = REG(INTENSITYINC);
-
+
gd_ia = gouraud_add & 0x00FFFFFF;
if (gd_ia & 0x00800000)
gd_ia = 0xFF000000 | gd_ia;
WriteLog(" GOURZ = %i\n",GOURZ);
WriteLog(" GOURD = %i\n",GOURD);
WriteLog(" SRCSHADE= %i\n",SRCSHADE);
- }
+ }
#endif
//NOTE: Pitch is ignored!
//Black is short by 3, pink is short by 1...
/*
Blit! (00110000 <- 000BF010) count: 9 x 31, A1/2_FLAGS: 000042E2/00010020 [cmd: 00010200]
- CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
+ CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A1 x/y: 262/124, A2 x/y: 128/0
Blit! (00110000 <- 000BF010) count: 5 x 38, A1/2_FLAGS: 000042E2/00010020 [cmd: 00010200]
- CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
+ CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A1 x/y: 264/117, A2 x/y: 407/0
Blit! (00110000 <- 000BF010) count: 9 x 23, A1/2_FLAGS: 000042E2/00010020 [cmd: 00010200]
- CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
+ CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
A1 step values: -10 (X), 1 (Y)
A1 -> pitch: 4(2) phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
A2 -> pitch: 1(0) phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
A1 x/y: 262/132, A2 x/y: 129/0
Blit! (00110000 <- 000BF010) count: 5 x 27, A1/2_FLAGS: 000042E2/00010020 [cmd: 00010200]
- CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
+ CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
A1 step values: -8 (X), 1 (Y)
A1 -> pitch: 4(2) phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
A2 -> pitch: 1(0) phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
; This looks like the ship icon in the upper left corner...
Blit! (00110000 <- 0010B2A8) count: 11 x 12, A1/2_FLAGS: 000042E2/00000020 [cmd: 09800609]
- CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
+ CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
A1 step values: -12 (X), 1 (Y)
A2 step values: 0 (X), 0 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
; D
Blit! (00110000 <- 0010B2A8) count: 12 x 12, A1/2_FLAGS: 000042E2/00000020 [cmd: 09800609]
- CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
+ CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
A1 step values: -14 (X), 1 (Y)
A2 step values: -4 (X), 0 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
; E
Blit! (00110000 <- 0010B2A8) count: 12 x 12, A1/2_FLAGS: 000042E2/00000020 [cmd: 09800609]
- CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
+ CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
A1 step values: -13 (X), 1 (Y)
A2 step values: -4 (X), 0 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
; M
Blit! (00110000 <- 0010B2A8) count: 12 x 12, A1/2_FLAGS: 000042E2/00000020 [cmd: 09800609]
- CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
+ CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
A1 step values: -12 (X), 1 (Y)
A2 step values: 0 (X), 0 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
; O
Blit! (00110000 <- 0010B2A8) count: 12 x 12, A1/2_FLAGS: 000042E2/00000020 [cmd: 09800609]
- CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
+ CMD -> src: SRCEN dst: DSTEN misc: a1ctl: UPDA1 UPDA2 mode: ity: z-op: op: LFU_REPLACE ctrl: DCOMPEN
A1 step values: -15 (X), 1 (Y)
A2 step values: -4 (X), 0 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
char * bppStr[8] = { "1bpp\0", "2bpp\0", "4bpp\0", "8bpp\0", "16bpp\0", "32bpp\0", "???\0", "!!!\0" };
char * opStr[16] = { "LFU_CLEAR", "LFU_NSAND", "LFU_NSAD", "LFU_NOTS", "LFU_SAND", "LFU_NOTD", "LFU_N_SXORD", "LFU_NSORND",
"LFU_SAD", "LFU_XOR", "LFU_D", "LFU_NSORD", "LFU_REPLACE", "LFU_SORND", "LFU_SORD", "LFU_ONE" };
- uint32 /*src = cmd & 0x07, dst = (cmd >> 3) & 0x07, misc = (cmd >> 6) & 0x03,
+ uint32_t /*src = cmd & 0x07, dst = (cmd >> 3) & 0x07, misc = (cmd >> 6) & 0x03,
a1ctl = (cmd >> 8) & 0x07,*/ mode = (cmd >> 11) & 0x07/*, ity = (cmd >> 14) & 0x0F,
zop = (cmd >> 18) & 0x07, op = (cmd >> 21) & 0x0F, ctrl = (cmd >> 25) & 0x3F*/;
- UINT32 a1f = REG(A1_FLAGS), a2f = REG(A2_FLAGS);
- uint32 p1 = a1f & 0x07, p2 = a2f & 0x07,
+ uint32_t a1f = REG(A1_FLAGS), a2f = REG(A2_FLAGS);
+ uint32_t p1 = a1f & 0x07, p2 = a2f & 0x07,
d1 = (a1f >> 3) & 0x07, d2 = (a2f >> 3) & 0x07,
zo1 = (a1f >> 6) & 0x07, zo2 = (a2f >> 6) & 0x07,
w1 = (a1f >> 9) & 0x3F, w2 = (a2f >> 9) & 0x3F,
ac1 = (a1f >> 16) & 0x1F, ac2 = (a2f >> 16) & 0x1F;
- UINT32 iw1 = ((0x04 | (w1 & 0x03)) << ((w1 & 0x3C) >> 2)) >> 2;
- UINT32 iw2 = ((0x04 | (w2 & 0x03)) << ((w2 & 0x3C) >> 2)) >> 2;
+ uint32_t iw1 = ((0x04 | (w1 & 0x03)) << ((w1 & 0x3C) >> 2)) >> 2;
+ uint32_t iw2 = ((0x04 | (w2 & 0x03)) << ((w2 & 0x3C) >> 2)) >> 2;
WriteLog("Blit! (%08X %s %08X) count: %d x %d, A1/2_FLAGS: %08X/%08X [cmd: %08X]\n", a1_addr, (mode&0x01 ? "->" : "<-"), a2_addr, n_pixels, n_lines, a1f, a2f, cmd);
// WriteLog(" CMD -> src: %d, dst: %d, misc: %d, a1ctl: %d, mode: %d, ity: %1X, z-op: %d, op: %1X, ctrl: %02X\n", src, dst, misc, a1ctl, mode, ity, zop, op, ctrl);
********************** STUFF CUT ABOVE THIS LINE! ******************************
*******************************************************************************/
-void blitter_init(void)
+void BlitterInit(void)
{
- blitter_reset();
+ BlitterReset();
}
-void blitter_reset(void)
+void BlitterReset(void)
{
memset(blitter_ram, 0x00, 0xA0);
}
-void blitter_done(void)
+void BlitterDone(void)
{
WriteLog("BLIT: Done.\n");
}
-uint8 BlitterReadByte(uint32 offset, uint32 who/*=UNKNOWN*/)
+uint8_t BlitterReadByte(uint32_t offset, uint32_t who/*=UNKNOWN*/)
{
offset &= 0xFF;
// status register
//This isn't cycle accurate--how to fix? !!! FIX !!!
//Probably have to do some multi-threaded implementation or at least a reentrant safe implementation...
+//Real hardware returns $00000805, just like the JTRM says.
+ if (offset == (0x38 + 0))
+ return 0x00;
+ if (offset == (0x38 + 1))
+ return 0x00;
+ if (offset == (0x38 + 2))
+ return 0x08;
if (offset == (0x38 + 3))
- return 0x01; // always idle
+ return 0x05; // always idle/never stopped (collision detection ignored!)
-// CHECK HERE ONCE THIS FIX HAS BEEN TESTED: [ ]
+// CHECK HERE ONCE THIS FIX HAS BEEN TESTED: [X]
//Fix for AvP:
if (offset >= 0x04 && offset <= 0x07)
//This is it. I wonder if it just ignores the lower three bits?
}
//Crappy!
-uint16 BlitterReadWord(uint32 offset, uint32 who/*=UNKNOWN*/)
+uint16_t BlitterReadWord(uint32_t offset, uint32_t who/*=UNKNOWN*/)
{
- return ((uint16)BlitterReadByte(offset, who) << 8) | (uint16)BlitterReadByte(offset+1, who);
+ return ((uint16_t)BlitterReadByte(offset, who) << 8) | (uint16_t)BlitterReadByte(offset+1, who);
}
//Crappy!
-uint32 BlitterReadLong(uint32 offset, uint32 who/*=UNKNOWN*/)
+uint32_t BlitterReadLong(uint32_t offset, uint32_t who/*=UNKNOWN*/)
{
return (BlitterReadWord(offset, who) << 16) | BlitterReadWord(offset+2, who);
}
-void BlitterWriteByte(uint32 offset, uint8 data, uint32 who/*=UNKNOWN*/)
+void BlitterWriteByte(uint32_t offset, uint8_t data, uint32_t who/*=UNKNOWN*/)
{
/*if (offset & 0xFF == 0x7B)
WriteLog("--> Wrote to B_STOP: value -> %02X\n", data);*/
case 0x81: blitter_ram[PATTERNDATA + 5] = data; break;
case 0x82: blitter_ram[SRCDATA + 4] = data; break;
case 0x83: blitter_ram[SRCDATA + 5] = data; break;
-
+
case 0x84: break;
case 0x85: blitter_ram[PATTERNDATA + 3] = data; break;
case 0x86: blitter_ram[SRCDATA + 2] = data; break;
case 0x87: blitter_ram[SRCDATA + 3] = data; break;
-
+
case 0x88: break;
case 0x89: blitter_ram[PATTERNDATA + 1] = data; break;
case 0x8A: blitter_ram[SRCDATA + 0] = data; break;
case 0x91: blitter_ram[SRCZINT + 5] = data; break;
case 0x92: blitter_ram[SRCZFRAC + 4] = data; break;
case 0x93: blitter_ram[SRCZFRAC + 5] = data; break;
-
+
case 0x94: blitter_ram[SRCZINT + 2] = data; break;
case 0x95: blitter_ram[SRCZINT + 3] = data; break;
case 0x96: blitter_ram[SRCZFRAC + 2] = data; break;
case 0x97: blitter_ram[SRCZFRAC + 3] = data; break;
-
+
case 0x98: blitter_ram[SRCZINT + 0] = data; break;
case 0x99: blitter_ram[SRCZINT + 1] = data; break;
case 0x9A: blitter_ram[SRCZFRAC + 0] = data; break;
blitter_ram[offset] = data;
}
-void BlitterWriteWord(uint32 offset, uint16 data, uint32 who/*=UNKNOWN*/)
+void BlitterWriteWord(uint32_t offset, uint16_t data, uint32_t who/*=UNKNOWN*/)
{
/*if (((offset & 0xFF) >= PATTERNDATA) && ((offset & 0xFF) < PATTERNDATA + 8))
{
}
//F02278,9,A,B
-void BlitterWriteLong(uint32 offset, uint32 data, uint32 who/*=UNKNOWN*/)
+void BlitterWriteLong(uint32_t offset, uint32_t data, uint32_t who/*=UNKNOWN*/)
{
/*if (((offset & 0xFF) >= PATTERNDATA) && ((offset & 0xFF) < PATTERNDATA + 8))
{
void LogBlit(void)
{
- char * opStr[16] = { "LFU_CLEAR", "LFU_NSAND", "LFU_NSAD", "LFU_NOTS", "LFU_SAND", "LFU_NOTD", "LFU_N_SXORD", "LFU_NSORND",
+ const char * opStr[16] = { "LFU_CLEAR", "LFU_NSAND", "LFU_NSAD", "LFU_NOTS", "LFU_SAND", "LFU_NOTD", "LFU_N_SXORD", "LFU_NSORND",
"LFU_SAD", "LFU_XOR", "LFU_D", "LFU_NSORD", "LFU_REPLACE", "LFU_SORND", "LFU_SORD", "LFU_ONE" };
- uint32 cmd = GET32(blitter_ram, 0x38);
- UINT32 m = (REG(A1_FLAGS) >> 9) & 0x03, e = (REG(A1_FLAGS) >> 11) & 0x0F;
- UINT32 a1_width = ((0x04 | m) << e) >> 2;
+ uint32_t cmd = GET32(blitter_ram, 0x38);
+ uint32_t m = (REG(A1_FLAGS) >> 9) & 0x03, e = (REG(A1_FLAGS) >> 11) & 0x0F;
+ uint32_t a1_width = ((0x04 | m) << e) >> 2;
m = (REG(A2_FLAGS) >> 9) & 0x03, e = (REG(A2_FLAGS) >> 11) & 0x0F;
- UINT32 a2_width = ((0x04 | m) << e) >> 2;
+ uint32_t a2_width = ((0x04 | m) << e) >> 2;
WriteLog("Blit!\n");
WriteLog(" COMMAND = %08X\n", cmd);
REG(A1_FLAGS) & 0x00003, (REG(A1_FLAGS) & 0x00038) >> 3,
(REG(A1_FLAGS) & 0x001C0) >> 6, a1_width, (REG(A1_FLAGS) & 0x30000) >> 16);
WriteLog(" a1_clip = %u, %u (%08X)\n", GET16(blitter_ram, A1_CLIP + 2), GET16(blitter_ram, A1_CLIP + 0), GET32(blitter_ram, A1_CLIP));
- WriteLog(" a1_pixel = %d, %d (%08X)\n", (int16)GET16(blitter_ram, A1_PIXEL + 2), (int16)GET16(blitter_ram, A1_PIXEL + 0), GET32(blitter_ram, A1_PIXEL));
- WriteLog(" a1_step = %d, %d (%08X)\n", (int16)GET16(blitter_ram, A1_STEP + 2), (int16)GET16(blitter_ram, A1_STEP + 0), GET32(blitter_ram, A1_STEP));
+ WriteLog(" a1_pixel = %d, %d (%08X)\n", (int16_t)GET16(blitter_ram, A1_PIXEL + 2), (int16_t)GET16(blitter_ram, A1_PIXEL + 0), GET32(blitter_ram, A1_PIXEL));
+ WriteLog(" a1_step = %d, %d (%08X)\n", (int16_t)GET16(blitter_ram, A1_STEP + 2), (int16_t)GET16(blitter_ram, A1_STEP + 0), GET32(blitter_ram, A1_STEP));
WriteLog(" a1_fstep = %u, %u (%08X)\n", GET16(blitter_ram, A1_FSTEP + 2), GET16(blitter_ram, A1_FSTEP + 0), GET32(blitter_ram, A1_FSTEP));
WriteLog(" a1_fpixel= %u, %u (%08X)\n", GET16(blitter_ram, A1_FPIXEL + 2), GET16(blitter_ram, A1_FPIXEL + 0), GET32(blitter_ram, A1_FPIXEL));
- WriteLog(" a1_inc = %d, %d (%08X)\n", (int16)GET16(blitter_ram, A1_INC + 2), (int16)GET16(blitter_ram, A1_INC + 0), GET32(blitter_ram, A1_INC));
+ WriteLog(" a1_inc = %d, %d (%08X)\n", (int16_t)GET16(blitter_ram, A1_INC + 2), (int16_t)GET16(blitter_ram, A1_INC + 0), GET32(blitter_ram, A1_INC));
WriteLog(" a1_finc = %u, %u (%08X)\n", GET16(blitter_ram, A1_FINC + 2), GET16(blitter_ram, A1_FINC + 0), GET32(blitter_ram, A1_FINC));
WriteLog(" a2_base = %08X\n", REG(A2_BASE));
REG(A2_FLAGS) & 0x00003, (REG(A2_FLAGS) & 0x00038) >> 3,
(REG(A2_FLAGS) & 0x001C0) >> 6, a2_width, (REG(A2_FLAGS) & 0x30000) >> 16);
WriteLog(" a2_mask = %u, %u (%08X)\n", GET16(blitter_ram, A2_MASK + 2), GET16(blitter_ram, A2_MASK + 0), GET32(blitter_ram, A2_MASK));
- WriteLog(" a2_pixel = %d, %d (%08X)\n", (int16)GET16(blitter_ram, A2_PIXEL + 2), (int16)GET16(blitter_ram, A2_PIXEL + 0), GET32(blitter_ram, A2_PIXEL));
- WriteLog(" a2_step = %d, %d (%08X)\n", (int16)GET16(blitter_ram, A2_STEP + 2), (int16)GET16(blitter_ram, A2_STEP + 0), GET32(blitter_ram, A2_STEP));
+ WriteLog(" a2_pixel = %d, %d (%08X)\n", (int16_t)GET16(blitter_ram, A2_PIXEL + 2), (int16_t)GET16(blitter_ram, A2_PIXEL + 0), GET32(blitter_ram, A2_PIXEL));
+ WriteLog(" a2_step = %d, %d (%08X)\n", (int16_t)GET16(blitter_ram, A2_STEP + 2), (int16_t)GET16(blitter_ram, A2_STEP + 0), GET32(blitter_ram, A2_STEP));
WriteLog(" count = %d x %d\n", GET16(blitter_ram, PIXLINECOUNTER + 2), GET16(blitter_ram, PIXLINECOUNTER));
WriteLog(" UPDA2 = %s\n", (UPDA2 ? "1" : "0"));
WriteLog(" DSTA2 = %s\n", (DSTA2 ? "1" : "0"));
WriteLog(" ZOP = %s %s %s\n", (Z_OP_INF ? "<" : ""), (Z_OP_EQU ? "=" : ""), (Z_OP_SUP ? ">" : ""));
- WriteLog("--LFUFUNC = %s\n", opStr[(cmd >> 21) & 0x0F]);
+ WriteLog("+-LFUFUNC = %s\n", opStr[(cmd >> 21) & 0x0F]);
WriteLog("| PATDSEL = %s (PD=%08X%08X)\n", (PATDSEL ? "1" : "0"), REG(PATTERNDATA), REG(PATTERNDATA + 4));
- WriteLog("--ADDDSEL = %s\n", (ADDDSEL ? "1" : "0"));
+ WriteLog("+-ADDDSEL = %s\n", (ADDDSEL ? "1" : "0"));
WriteLog(" CMPDST = %s\n", (CMPDST ? "1" : "0"));
WriteLog(" BCOMPEN = %s\n", (BCOMPEN ? "1" : "0"));
WriteLog(" DCOMPEN = %s\n", (DCOMPEN ? "1" : "0"));
#define TXTEXT (false)
#define POLYGON (false)
-void BlitterMidsummer(uint32 cmd)
+void BlitterMidsummer(uint32_t cmd)
{
-uint32 outer_loop, inner_loop, a1_addr, a2_addr;
-int32 a1_x, a1_y, a2_x, a2_y, a1_width, a2_width;
-uint8 a1_phrase_mode, a2_phrase_mode;
+#ifdef LOG_BLITS
+ LogBlit();
+#endif
+uint32_t outer_loop, inner_loop, a1_addr, a2_addr;
+int32_t a1_x, a1_y, a2_x, a2_y, a1_width, a2_width;
+uint8_t a1_phrase_mode, a2_phrase_mode;
a1_addr = REG(A1_BASE) & 0xFFFFFFF8;
a2_addr = REG(A2_BASE) & 0xFFFFFFF8;
a1_x = (REG(A1_PIXEL) << 16) | (REG(A1_FPIXEL) & 0xFFFF);
a1_y = (REG(A1_PIXEL) & 0xFFFF0000) | (REG(A1_FPIXEL) >> 16);
- UINT32 m = (REG(A1_FLAGS) >> 9) & 0x03, e = (REG(A1_FLAGS) >> 11) & 0x0F;
+ uint32_t m = (REG(A1_FLAGS) >> 9) & 0x03, e = (REG(A1_FLAGS) >> 11) & 0x0F;
a1_width = ((0x04 | m) << e) >> 2;//*/
a2_x = (REG(A2_PIXEL) & 0x0000FFFF) << 16;
a2_y = (REG(A2_PIXEL) & 0xFFFF0000);
LogBlit();
}//*/
- uint64 srcData = GET64(blitter_ram, SRCDATA), srcXtraData,
+ uint64_t srcData = GET64(blitter_ram, SRCDATA), srcXtraData,
dstData = GET64(blitter_ram, DSTDATA), writeData;
- uint32 srcAddr, dstAddr;
- uint8 bitCount, a1PixelSize, a2PixelSize;
+ uint32_t srcAddr, dstAddr;
+ uint8_t bitCount, a1PixelSize, a2PixelSize;
// JTRM says phrase mode only works for 8BPP or higher, so let's try this...
- uint32 phraseOffset[8] = { 8, 8, 8, 8, 4, 2, 0, 0 };
- uint8 pixelShift[8] = { 3, 2, 1, 0, 1, 2, 0, 0 };
+ uint32_t phraseOffset[8] = { 8, 8, 8, 8, 4, 2, 0, 0 };
+ uint8_t pixelShift[8] = { 3, 2, 1, 0, 1, 2, 0, 0 };
a1PixelSize = (blitter_ram[A1_FLAGS + 3] >> 3) & 0x07;
a2PixelSize = (blitter_ram[A2_FLAGS + 3] >> 3) & 0x07;
//pointing at. Likewise, the pixel (if in BPP 1, 2 & 4, chopped) otherwise. It probably still
//transfers an entire phrase even in pixel mode.
//Odd thought: Does it expand, e.g., 1 BPP pixels into 32 BPP internally? Hmm...
-//No.
+//No.
/*
a1_addr = REG(A1_BASE) & 0xFFFFFFF8;
a2_addr = REG(A2_BASE) & 0xFFFFFFF8;
a1_width = ((0x04 | m) << e) >> 2;
a2_width = ((0x04 | m) << e) >> 2;
- // write values back to registers
+ // write values back to registers
WREG(A1_PIXEL, (a1_y & 0xFFFF0000) | ((a1_x >> 16) & 0xFFFF));
WREG(A1_FPIXEL, (a1_y << 16) | (a1_x & 0xFFFF));
WREG(A2_PIXEL, (a2_y & 0xFFFF0000) | ((a2_x >> 16) & 0xFFFF));
}
else*/
{
-// uint32 pixAddr = ((DSTA2 ? a1_x : a2_x) >> 16)
+// uint32_t pixAddr = ((DSTA2 ? a1_x : a2_x) >> 16)
// + (((DSTA2 ? a1_y : a2_y) >> 16) * (DSTA2 ? a1_width : a2_width));
- int32 pixAddr = (int16)((DSTA2 ? a1_x : a2_x) >> 16)
- + ((int16)((DSTA2 ? a1_y : a2_y) >> 16) * (DSTA2 ? a1_width : a2_width));
+ int32_t pixAddr = (int16_t)((DSTA2 ? a1_x : a2_x) >> 16)
+ + ((int16_t)((DSTA2 ? a1_y : a2_y) >> 16) * (DSTA2 ? a1_width : a2_width));
if ((DSTA2 ? a1PixelSize : a2PixelSize) < 3)
pixAddr >>= pixelShift[(DSTA2 ? a1PixelSize : a2PixelSize)];
if ((DSTA2 ? a1_phrase_mode : a2_phrase_mode) == 1)
{
- srcData = ((uint64)JaguarReadLong(srcAddr, BLITTER) << 32)
- | (uint64)JaguarReadLong(srcAddr + 4, BLITTER);
+ srcData = ((uint64_t)JaguarReadLong(srcAddr, BLITTER) << 32)
+ | (uint64_t)JaguarReadLong(srcAddr + 4, BLITTER);
}
else
{
#ifdef LOG_BLITTER_MEMORY_ACCESSES
if (logBlit)
- WriteLog("BLITTER: srcAddr=%08X, srcData=%08X %08X\n", srcAddr, (uint32)(srcData >> 32), (uint32)(srcData & 0xFFFFFFFF));
+ WriteLog("BLITTER: srcAddr=%08X, srcData=%08X %08X\n", srcAddr, (uint32_t)(srcData >> 32), (uint32_t)(srcData & 0xFFFFFFFF));
#endif
if (SRCENZ)
dstAddr = (DSTA2 ? a2_addr : a1_addr);
{
-// uint32 pixAddr = ((DSTA2 ? a2_x : a1_x) >> 16)
+// uint32_t pixAddr = ((DSTA2 ? a2_x : a1_x) >> 16)
// + (((DSTA2 ? a2_y : a1_y) >> 16) * (DSTA2 ? a2_width : a1_width));
- int32 pixAddr = (int16)((DSTA2 ? a2_x : a1_x) >> 16)
- + ((int16)((DSTA2 ? a2_y : a1_y) >> 16) * (DSTA2 ? a2_width : a1_width));
+ int32_t pixAddr = (int16_t)((DSTA2 ? a2_x : a1_x) >> 16)
+ + ((int16_t)((DSTA2 ? a2_y : a1_y) >> 16) * (DSTA2 ? a2_width : a1_width));
if ((DSTA2 ? a2PixelSize : a1PixelSize) < 3)
pixAddr >>= pixelShift[(DSTA2 ? a2PixelSize : a1PixelSize)];
if ((DSTA2 ? a2_phrase_mode : a1_phrase_mode) == 1)
{
- dstData = ((uint64)JaguarReadLong(srcAddr, BLITTER) << 32)
- | (uint64)JaguarReadLong(srcAddr + 4, BLITTER);
+ dstData = ((uint64_t)JaguarReadLong(srcAddr, BLITTER) << 32)
+ | (uint64_t)JaguarReadLong(srcAddr + 4, BLITTER);
}
else
{
#ifdef LOG_BLITTER_MEMORY_ACCESSES
if (logBlit)
- WriteLog("BLITTER (dread): dstAddr=%08X, dstData=%08X %08X\n", dstAddr, (uint32)(dstData >> 32), (uint32)(dstData & 0xFFFFFFFF));
+ WriteLog("BLITTER (dread): dstAddr=%08X, dstData=%08X %08X\n", dstAddr, (uint32_t)(dstData >> 32), (uint32_t)(dstData & 0xFFFFFFFF));
#endif
if (DSTENZ)
if (CLIPA1)
{
- uint16 x = a1_x >> 16, y = a1_y >> 16;
+ uint16_t x = a1_x >> 16, y = a1_y >> 16;
if (x >= GET16(blitter_ram, A1_CLIP + 2) || y >= GET16(blitter_ram, A1_CLIP))
goto inhibitWrite;
}
// Figure out what gets written...
-
+
if (PATDSEL)
{
writeData = GET64(blitter_ram, PATTERNDATA);
else // LFUFUNC is the default...
{
writeData = 0;
-
+
if (LFU_NAN)
writeData |= ~srcData & ~dstData;
if (LFU_NA)
{
/* dstAddr += ((DSTA2 ? a2_x : a1_x) >> 16)
+ (((DSTA2 ? a2_y : a1_y) >> 16) * (DSTA2 ? a2_width : a1_width));*/
-// uint32 pixAddr = ((DSTA2 ? a2_x : a1_x) >> 16)
+// uint32_t pixAddr = ((DSTA2 ? a2_x : a1_x) >> 16)
// + (((DSTA2 ? a2_y : a1_y) >> 16) * (DSTA2 ? a2_width : a1_width));
- int32 pixAddr = (int16)((DSTA2 ? a2_x : a1_x) >> 16)
- + ((int16)((DSTA2 ? a2_y : a1_y) >> 16) * (DSTA2 ? a2_width : a1_width));
+ int32_t pixAddr = (int16_t)((DSTA2 ? a2_x : a1_x) >> 16)
+ + ((int16_t)((DSTA2 ? a2_y : a1_y) >> 16) * (DSTA2 ? a2_width : a1_width));
if ((DSTA2 ? a2PixelSize : a1PixelSize) < 3)
pixAddr >>= pixelShift[(DSTA2 ? a2PixelSize : a1PixelSize)];
#ifdef LOG_BLITTER_MEMORY_ACCESSES
if (logBlit)
- WriteLog("BLITTER: dstAddr=%08X, writeData=%08X %08X\n", dstAddr, (uint32)(writeData >> 32), (uint32)(writeData & 0xFFFFFFFF));
+ WriteLog("BLITTER: dstAddr=%08X, writeData=%08X %08X\n", dstAddr, (uint32_t)(writeData >> 32), (uint32_t)(writeData & 0xFFFFFFFF));
#endif
inhibitWrite://Should this go here? or on the other side of the X/Y incrementing?
a2_x += (blitter_ram[A2_FLAGS + 1] & 0x08 ? -1 << 16 : 1 << 16);
/* else if ((blitter_ram[A2_FLAGS + 1] & 0x03) == 2)
a2_x += 0 << 16; */
-
+
if (blitter_ram[A2_FLAGS + 1] & 0x04)
a2_y += (blitter_ram[A2_FLAGS + 1] & 0x10 ? -1 << 16 : 1 << 16);
goto a1update
*/
/*
-#define A1_PIXEL ((UINT32)0x0C) // Integer part of the pixel (Y.i and X.i)
-#define A1_STEP ((UINT32)0x10) // Integer part of the step
-#define A1_FSTEP ((UINT32)0x14) // Fractional part of the step
-#define A1_FPIXEL ((UINT32)0x18) // Fractional part of the pixel (Y.f and X.f)
+#define A1_PIXEL ((uint32_t)0x0C) // Integer part of the pixel (Y.i and X.i)
+#define A1_STEP ((uint32_t)0x10) // Integer part of the step
+#define A1_FSTEP ((uint32_t)0x14) // Fractional part of the step
+#define A1_FPIXEL ((uint32_t)0x18) // Fractional part of the pixel (Y.f and X.f)
*/
// This is all kinda murky. All we have are the Midsummer docs to give us any guidance,
// and it's incomplete or filled with errors (like above). Aarrrgggghhhhh!
//This isn't right. Is it? I don't think the fractional parts are signed...
-// a1_x += (int32)((int16)GET16(blitter_ram, A1_FSTEP + 2));
-// a1_y += (int32)((int16)GET16(blitter_ram, A1_FSTEP + 0));
+// a1_x += (int32_t)((int16_t)GET16(blitter_ram, A1_FSTEP + 2));
+// a1_y += (int32_t)((int16_t)GET16(blitter_ram, A1_FSTEP + 0));
a1_x += GET16(blitter_ram, A1_FSTEP + 2);
a1_y += GET16(blitter_ram, A1_FSTEP + 0);
else if DATINIT goto init_if
else restart inner
*/
- a1_x += (int32)(GET16(blitter_ram, A1_STEP + 2) << 16);
- a1_y += (int32)(GET16(blitter_ram, A1_STEP + 0) << 16);
+ a1_x += (int32_t)(GET16(blitter_ram, A1_STEP + 2) << 16);
+ a1_y += (int32_t)(GET16(blitter_ram, A1_STEP + 0) << 16);
//kill this, for now...
if DATINIT goto init_if
else restart inner
*/
- a2_x += (int32)(GET16(blitter_ram, A2_STEP + 2) << 16);
- a2_y += (int32)(GET16(blitter_ram, A2_STEP + 0) << 16);
+ a2_x += (int32_t)(GET16(blitter_ram, A2_STEP + 2) << 16);
+ a2_y += (int32_t)(GET16(blitter_ram, A2_STEP + 0) << 16);
if (DATINIT)
}
#endif
-/*
-Here's a collection of various blits that aren't handled correctly yet...
-
-Spinning cube in Jaguar BIOS opening:
-
- COMMAND = 41802F41
- a1_base = 0015B000
- a1_flags = 00033020 (0 0 0 11 . 011000 000 100 . 00)
- pitch=0, pixSz=4, zOff=0, width=64, xCtrl=3
- a1_clip = 64, 64 (00400040)
- a1_pixel = 0, 65534 (FFFE0000)
- a1_step = -64, -1 (FFFFFFC0)
- a1_fstep = 0, 18161 (46F10000)
- a1_fpixel= 32768, 42497 (A6018000)
- a1_inc = 1, 0 (00000001)
- a1_finc = 0, 1926 (07860000)
- a2_base = 0014B000
- a2_flags = 00013820 (0 0 0 01 0 011100 000 100 . 00)
- pitch=0, pixSz=4, zOff=0, width=128, xCtrl=1
- a2_mask = 0, 0 (00000000)
- a2_pixel = 33, 55 (00370021)
- a2_step = -64, 1 (0001FFC0)
- count = 64 x 57
- SRCEN = 1
- SRCENZ = 0
- SRCENX = 0
- DSTEN = 0
- DSTENZ = 0
- DSTWRZ = 0
- CLIPA1 = 1
- UPDA1F = 1
- UPDA1 = 1
- UPDA2 = 1
- DSTA2 = 1
- ZOP =
---LFUFUNC = LFU_REPLACE
-| PATDSEL = 0 (PD=0000000000000000)
---ADDDSEL = 0
- CMPDST = 0
- BCOMPEN = 0
- DCOMPEN = 0
- TOPBEN = 0
- TOPNEN = 0
- BKGWREN = 0
- GOURD = 0 (II=00FCDC80, SD=FF00FF00FF00FF00)
- GOURZ = 1 (ZI=00000000, ZD=0000000000000000, SZ1=0000000000000000, SZ2=0000000000000000)
- SRCSHADE = 1
-*/
+//
+// Here's attempt #2--taken from the Oberon chip specs!
+//
#ifdef USE_MIDSUMMER_BLITTER_MKII
-// Here's attempt #2--taken from the Oberon chip specs!
-void ADDRGEN(uint32 &, uint32 &, bool, bool,
- uint16, uint16, uint32, uint8, uint8, uint8, uint8,
- uint16, uint16, uint32, uint8, uint8, uint8, uint8);
-void ADDARRAY(uint16 * addq, uint8 daddasel, uint8 daddbsel, uint8 daddmode,
- uint64 dstd, uint32 iinc, uint8 initcin[], uint64 initinc, uint16 initpix,
- uint32 istep, uint64 patd, uint64 srcd, uint64 srcz1, uint64 srcz2,
- uint32 zinc, uint32 zstep);
-void ADD16SAT(uint16 &r, uint8 &co, uint16 a, uint16 b, uint8 cin, bool sat, bool eightbit, bool hicinh);
-void ADDAMUX(int16 &adda_x, int16 &adda_y, uint8 addasel, int16 a1_step_x, int16 a1_step_y,
- int16 a1_stepf_x, int16 a1_stepf_y, int16 a2_step_x, int16 a2_step_y,
- int16 a1_inc_x, int16 a1_inc_y, int16 a1_incf_x, int16 a1_incf_y, uint8 adda_xconst,
+void ADDRGEN(uint32_t &, uint32_t &, bool, bool,
+ uint16_t, uint16_t, uint32_t, uint8_t, uint8_t, uint8_t, uint8_t,
+ uint16_t, uint16_t, uint32_t, uint8_t, uint8_t, uint8_t, uint8_t);
+void ADDARRAY(uint16_t * addq, uint8_t daddasel, uint8_t daddbsel, uint8_t daddmode,
+ uint64_t dstd, uint32_t iinc, uint8_t initcin[], uint64_t initinc, uint16_t initpix,
+ uint32_t istep, uint64_t patd, uint64_t srcd, uint64_t srcz1, uint64_t srcz2,
+ uint32_t zinc, uint32_t zstep);
+void ADD16SAT(uint16_t &r, uint8_t &co, uint16_t a, uint16_t b, uint8_t cin, bool sat, bool eightbit, bool hicinh);
+void ADDAMUX(int16_t &adda_x, int16_t &adda_y, uint8_t addasel, int16_t a1_step_x, int16_t a1_step_y,
+ int16_t a1_stepf_x, int16_t a1_stepf_y, int16_t a2_step_x, int16_t a2_step_y,
+ int16_t a1_inc_x, int16_t a1_inc_y, int16_t a1_incf_x, int16_t a1_incf_y, uint8_t adda_xconst,
bool adda_yconst, bool addareg, bool suba_x, bool suba_y);
-void ADDBMUX(int16 &addb_x, int16 &addb_y, uint8 addbsel, int16 a1_x, int16 a1_y,
- int16 a2_x, int16 a2_y, int16 a1_frac_x, int16 a1_frac_y);
-void DATAMUX(int16 &data_x, int16 &data_y, uint32 gpu_din, int16 addq_x, int16 addq_y, bool addqsel);
-void ADDRADD(int16 &addq_x, int16 &addq_y, bool a1fracldi,
- uint16 adda_x, uint16 adda_y, uint16 addb_x, uint16 addb_y, uint8 modx, bool suba_x, bool suba_y);
-void DATA (uint64 &wdata, uint8 &dcomp, uint8 &zcomp, bool &nowrite,
- bool big_pix, bool cmpdst, uint8 daddasel, uint8 daddbsel, uint8 daddmode, bool daddq_sel, uint8 data_sel,
- uint8 dbinh, uint8 dend, uint8 dstart, uint64 dstd, uint32 iinc, uint8 lfu_func, uint64 patd, bool patdadd,
- bool phrase_mode, uint64 srcd, bool srcdread, bool srczread, bool srcz2add, uint8 zmode,
- bool bcompen, bool bkgwren, bool dcompen, uint8 icount, uint8 pixsize);
-void COMP_CTRL(uint8 &dbinh, bool &nowrite,
- bool bcompen, bool big_pix, bool bkgwren, uint8 dcomp, bool dcompen, uint8 icount,
- uint8 pixsize, bool phrase_mode, uint8 srcd, uint8 zcomp);
+void ADDBMUX(int16_t &addb_x, int16_t &addb_y, uint8_t addbsel, int16_t a1_x, int16_t a1_y,
+ int16_t a2_x, int16_t a2_y, int16_t a1_frac_x, int16_t a1_frac_y);
+void DATAMUX(int16_t &data_x, int16_t &data_y, uint32_t gpu_din, int16_t addq_x, int16_t addq_y, bool addqsel);
+void ADDRADD(int16_t &addq_x, int16_t &addq_y, bool a1fracldi,
+ uint16_t adda_x, uint16_t adda_y, uint16_t addb_x, uint16_t addb_y, uint8_t modx, bool suba_x, bool suba_y);
+void DATA(uint64_t &wdata, uint8_t &dcomp, uint8_t &zcomp, bool &nowrite,
+ bool big_pix, bool cmpdst, uint8_t daddasel, uint8_t daddbsel, uint8_t daddmode, bool daddq_sel, uint8_t data_sel,
+ uint8_t dbinh, uint8_t dend, uint8_t dstart, uint64_t dstd, uint32_t iinc, uint8_t lfu_func, uint64_t &patd, bool patdadd,
+ bool phrase_mode, uint64_t srcd, bool srcdread, bool srczread, bool srcz2add, uint8_t zmode,
+ bool bcompen, bool bkgwren, bool dcompen, uint8_t icount, uint8_t pixsize,
+ uint64_t &srcz, uint64_t dstz, uint32_t zinc);
+void COMP_CTRL(uint8_t &dbinh, bool &nowrite,
+ bool bcompen, bool big_pix, bool bkgwren, uint8_t dcomp, bool dcompen, uint8_t icount,
+ uint8_t pixsize, bool phrase_mode, uint8_t srcd, uint8_t zcomp);
#define VERBOSE_BLITTER_LOGGING
bool logBlit = false;
void BlitterMidsummer2(void)
{
+#ifdef LOG_BLITS
+ LogBlit();
+#endif
// Here's what the specs say the state machine does. Note that this can probably be
// greatly simplified (also, it's different from what John has in his Oberon docs):
//Will remove stuff that isn't in Jaguar I once fully described (stuff like texture won't
//be described here at all)...
- uint32 cmd = GET32(blitter_ram, COMMAND);
+ uint32_t cmd = GET32(blitter_ram, COMMAND);
-// $01800005 has SRCENX, may have to investigate further...
-// $00011008 has GOURD & DSTEN.
-// $41802F41 has SRCSHADE, CLIPA1
logBlit = false;
-/*if (
- cmd != 0x00010200
- && cmd != 0x01800001
+if (
+ cmd != 0x00010200 && // PATDSEL
+ cmd != 0x01800001 // SRCEN LFUFUNC=C
&& cmd != 0x01800005
//Boot ROM ATARI letters:
&& cmd != 0x00011008 // DSTEN GOURD PATDSEL
//T2K TEMPEST letters:
&& cmd != 0x09800741 // SRCEN CLIP_A1 UPDA1 UPDA1F UPDA2 LFUFUNC=C DCOMPEN
//Static letters on Cybermorph intro screen:
- && cmd != 0x09800609 // DCOMPEN
+ && cmd != 0x09800609 // SRCEN DSTEN UPDA1 UPDA2 LFUFUNC=C DCOMPEN
//Static pic on title screen:
&& cmd != 0x01800601 // SRCEN UPDA1 UPDA2 LFUFUNC=C
//Turning letters on Cybermorph intro screen:
- && cmd != 0x09800F41 // SRCEN CLIP_A1 UPDA1 UPDA1F UPDA2 DSTA2 LFUFUNC=C DCOMPEN
+// && cmd != 0x09800F41 // SRCEN CLIP_A1 UPDA1 UPDA1F UPDA2 DSTA2 LFUFUNC=C DCOMPEN
&& cmd != 0x00113078 // DSTEN DSTENZ DSTWRZ CLIP_A1 GOURD GOURZ PATDSEL ZMODE=4
&& cmd != 0x09900F39 // SRCEN DSTEN DSTENZ DSTWRZ UPDA1 UPDA1F UPDA2 DSTA2 ZMODE=4 LFUFUNC=C DCOMPEN
&& cmd != 0x09800209 // SRCEN DSTEN UPDA1 LFUFUNC=C DCOMPEN
&& cmd != 0x00010000 // PATDSEL
//Hover Strike text:
&& cmd != 0x1401060C // SRCENX DSTEN UPDA1 UPDA2 PATDSEL BCOMPEN BKGWREN
+//Hover Strike 3D stuff
+ && cmd != 0x01902839 // SRCEN DSTEN DSTENZ DSTWRZ DSTA2 GOURZ ZMODE=4 LFUFUNC=C
+//Hover Strike darkening on intro to play (briefing) screen
+ && cmd != 0x00020208 // DSTEN UPDA1 ADDDSEL
+//Trevor McFur stuff:
+ && cmd != 0x05810601 // SRCEN UPDA1 UPDA2 PATDSEL BCOMPEN
+ && cmd != 0x01800201 // SRCEN UPDA1 LFUFUNC=C
+//T2K:
+ && cmd != 0x00011000 // GOURD PATDSEL
+ && cmd != 0x00011040 // CLIP_A1 GOURD PATDSEL
+//Checkered flag:
+ && cmd != 0x01800000 // LFUFUNC=C
+ && cmd != 0x01800401 //
+ && cmd != 0x01800040 //
+ && cmd != 0x00020008 //
+// && cmd != 0x09800F41 // SRCEN CLIP_A1 UPDA1 UPDA1F UPDA2 DSTA2 LFUFUNC=C DCOMPEN
)
logBlit = true;//*/
+//logBlit = true;
+if (blit_start_log == 0) // Wait for the signal...
+ logBlit = false;//*/
+/*
+Some T2K unique blits:
+logBlit = F, cmd = 00010200 *
+logBlit = F, cmd = 00011000
+logBlit = F, cmd = 00011040
+logBlit = F, cmd = 01800005 *
+logBlit = F, cmd = 09800741 *
+
+Hover Strike mission selection screen:
+Blit! (CMD = 01902839) // SRCEN DSTEN DSTENZ DSTWRZ DSTA2 GOURZ ZMODE=4 LFUFUNC=C
+
+Checkered Flag blits in the screw up zone:
+Blit! (CMD = 01800001) // SRCEN LFUFUNC=C
+Blit! (CMD = 01800000) // LFUFUNC=C
+Blit! (CMD = 00010000) // PATDSEL
+
+Wolfenstein 3D in the fuckup zone:
+Blit! (CMD = 01800000) // LFUFUNC=C
+*/
//printf("logBlit = %s, cmd = %08X\n", (logBlit ? "T" : "F"), cmd);
//fflush(stdout);
+//logBlit = true;
+
+/*
+Blit! (CMD = 00011040)
+Flags: CLIP_A1 GOURD PATDSEL
+ count = 18 x 1
+ a1_base = 00100000, a2_base = 0081F6A8
+ a1_x = 00A7, a1_y = 0014, a1_frac_x = 0000, a1_frac_y = 0000, a2_x = 0001, a2_y = 0000
+ a1_step_x = FE80, a1_step_y = 0001, a1_stepf_x = 0000, a1_stepf_y = 0000, a2_step_x = FFF8, a2_step_y = 0001
+ a1_inc_x = 0001, a1_inc_y = 0000, a1_incf_x = 0000, a1_incf_y = 0000
+ a1_win_x = 0180, a1_win_y = 0118, a2_mask_x = 0000, a2_mask_y = 0000
+ a2_mask=F a1add=+phr/+0 a2add=+phr/+0
+ a1_pixsize = 4, a2_pixsize = 4
+*/
+//Testing T2K...
+/*logBlit = false;
+if (cmd == 0x00011040
+ && (GET16(blitter_ram, A1_PIXEL + 2) == 0x00A7) && (GET16(blitter_ram, A1_PIXEL + 0) == 0x0014)
+ && (GET16(blitter_ram, A2_PIXEL + 2) == 0x0001) && (GET16(blitter_ram, A2_PIXEL + 0) == 0x0000)
+ && (GET16(blitter_ram, PIXLINECOUNTER + 2) == 18))
+ logBlit = true;*/
// Line states passed in via the command register
patdsel = (PATDSEL), adddsel = (ADDDSEL), cmpdst = (CMPDST), bcompen = (BCOMPEN),
dcompen = (DCOMPEN), bkgwren = (BKGWREN), srcshade = (SRCSHADE);
- uint8 zmode = (cmd & 0x01C0000) >> 18, lfufunc = (cmd & 0x1E00000) >> 21;
-//Missing: ZMODE, LFUFUNC, BUSHI
+ uint8_t zmode = (cmd & 0x01C0000) >> 18, lfufunc = (cmd & 0x1E00000) >> 21;
+//Missing: BUSHI
//Where to find various lines:
// clip_a1 -> inner
// gourd -> dcontrol, inner, outer, state
#endif
// Lines that don't exist in Jaguar I (and will never be asserted)
-
+
bool polygon = false, datinit = false, a1_stepld = false, a2_stepld = false, ext_int = false;
bool istepadd = false, istepfadd = false, finneradd = false, inneradd = false;
bool zstepfadd = false, zstepadd = false;
// Various registers set up by user
- uint16 ocount = GET16(blitter_ram, PIXLINECOUNTER);
- uint8 a1_pitch = blitter_ram[A1_FLAGS + 3] & 0x03;
- uint8 a2_pitch = blitter_ram[A2_FLAGS + 3] & 0x03;
- uint8 a1_pixsize = (blitter_ram[A1_FLAGS + 3] & 0x38) >> 3;
- uint8 a2_pixsize = (blitter_ram[A2_FLAGS + 3] & 0x38) >> 3;
- uint8 a1_zoffset = (GET16(blitter_ram, A1_FLAGS + 2) >> 6) & 0x07;
- uint8 a2_zoffset = (GET16(blitter_ram, A2_FLAGS + 2) >> 6) & 0x07;
- uint8 a1_width = (blitter_ram[A1_FLAGS + 2] >> 1) & 0x3F;
- uint8 a2_width = (blitter_ram[A2_FLAGS + 2] >> 1) & 0x3F;
+ uint16_t ocount = GET16(blitter_ram, PIXLINECOUNTER);
+ uint8_t a1_pitch = blitter_ram[A1_FLAGS + 3] & 0x03;
+ uint8_t a2_pitch = blitter_ram[A2_FLAGS + 3] & 0x03;
+ uint8_t a1_pixsize = (blitter_ram[A1_FLAGS + 3] & 0x38) >> 3;
+ uint8_t a2_pixsize = (blitter_ram[A2_FLAGS + 3] & 0x38) >> 3;
+ uint8_t a1_zoffset = (GET16(blitter_ram, A1_FLAGS + 2) >> 6) & 0x07;
+ uint8_t a2_zoffset = (GET16(blitter_ram, A2_FLAGS + 2) >> 6) & 0x07;
+ uint8_t a1_width = (blitter_ram[A1_FLAGS + 2] >> 1) & 0x3F;
+ uint8_t a2_width = (blitter_ram[A2_FLAGS + 2] >> 1) & 0x3F;
bool a2_mask = blitter_ram[A2_FLAGS + 2] & 0x80;
- uint8 a1addx = blitter_ram[A1_FLAGS + 1] & 0x03, a2addx = blitter_ram[A2_FLAGS + 1] & 0x03;
+ uint8_t a1addx = blitter_ram[A1_FLAGS + 1] & 0x03, a2addx = blitter_ram[A2_FLAGS + 1] & 0x03;
bool a1addy = blitter_ram[A1_FLAGS + 1] & 0x04, a2addy = blitter_ram[A2_FLAGS + 1] & 0x04;
bool a1xsign = blitter_ram[A1_FLAGS + 1] & 0x08, a2xsign = blitter_ram[A2_FLAGS + 1] & 0x08;
bool a1ysign = blitter_ram[A1_FLAGS + 1] & 0x10, a2ysign = blitter_ram[A2_FLAGS + 1] & 0x10;
- uint32 a1_base = GET32(blitter_ram, A1_BASE) & 0xFFFFFFF8; // Phrase aligned by ignoring bottom 3 bits
- uint32 a2_base = GET32(blitter_ram, A2_BASE) & 0xFFFFFFF8;
-
- uint16 a1_win_x = GET16(blitter_ram, A1_CLIP + 2) & 0x7FFF;
- uint16 a1_win_y = GET16(blitter_ram, A1_CLIP + 0) & 0x7FFF;
- int16 a1_x = (int16)GET16(blitter_ram, A1_PIXEL + 2);
- int16 a1_y = (int16)GET16(blitter_ram, A1_PIXEL + 0);
- int16 a1_step_x = (int16)GET16(blitter_ram, A1_STEP + 2);
- int16 a1_step_y = (int16)GET16(blitter_ram, A1_STEP + 0);
- uint16 a1_stepf_x = GET16(blitter_ram, A1_FSTEP + 2);
- uint16 a1_stepf_y = GET16(blitter_ram, A1_FSTEP + 0);
- uint16 a1_frac_x = GET16(blitter_ram, A1_FPIXEL + 2);
- uint16 a1_frac_y = GET16(blitter_ram, A1_FPIXEL + 0);
- int16 a1_inc_x = (int16)GET16(blitter_ram, A1_INC + 2);
- int16 a1_inc_y = (int16)GET16(blitter_ram, A1_INC + 0);
- uint16 a1_incf_x = GET16(blitter_ram, A1_FINC + 2);
- uint16 a1_incf_y = GET16(blitter_ram, A1_FINC + 0);
-
- int16 a2_x = (int16)GET16(blitter_ram, A2_PIXEL + 2);
- int16 a2_y = (int16)GET16(blitter_ram, A2_PIXEL + 0);
- uint16 a2_mask_x = GET16(blitter_ram, A2_MASK + 2);
- uint16 a2_mask_y = GET16(blitter_ram, A2_MASK + 0);
- int16 a2_step_x = (int16)GET16(blitter_ram, A2_STEP + 2);
- int16 a2_step_y = (int16)GET16(blitter_ram, A2_STEP + 0);
-
- uint64 srcd1 = GET64(blitter_ram, SRCDATA);
- uint64 srcd2 = 0;
- uint64 dstd = GET64(blitter_ram, DSTDATA);
- uint64 patd = GET64(blitter_ram, PATTERNDATA);
- uint32 iinc = GET32(blitter_ram, INTENSITYINC);
- uint64 srcz1 = GET64(blitter_ram, SRCZINT);
- uint64 srcz2 = GET64(blitter_ram, SRCZFRAC);
- uint64 dstz = GET64(blitter_ram, DSTZ);
- uint32 zinc = GET32(blitter_ram, ZINC);
- uint32 collision = GET32(blitter_ram, COLLISIONCTRL);// 0=RESUME, 1=ABORT, 2=STOPEN
-
- uint8 pixsize = (dsta2 ? a2_pixsize : a1_pixsize); // From ACONTROL
-
+ uint32_t a1_base = GET32(blitter_ram, A1_BASE) & 0xFFFFFFF8; // Phrase aligned by ignoring bottom 3 bits
+ uint32_t a2_base = GET32(blitter_ram, A2_BASE) & 0xFFFFFFF8;
+
+ uint16_t a1_win_x = GET16(blitter_ram, A1_CLIP + 2) & 0x7FFF;
+ uint16_t a1_win_y = GET16(blitter_ram, A1_CLIP + 0) & 0x7FFF;
+ int16_t a1_x = (int16_t)GET16(blitter_ram, A1_PIXEL + 2);
+ int16_t a1_y = (int16_t)GET16(blitter_ram, A1_PIXEL + 0);
+ int16_t a1_step_x = (int16_t)GET16(blitter_ram, A1_STEP + 2);
+ int16_t a1_step_y = (int16_t)GET16(blitter_ram, A1_STEP + 0);
+ uint16_t a1_stepf_x = GET16(blitter_ram, A1_FSTEP + 2);
+ uint16_t a1_stepf_y = GET16(blitter_ram, A1_FSTEP + 0);
+ uint16_t a1_frac_x = GET16(blitter_ram, A1_FPIXEL + 2);
+ uint16_t a1_frac_y = GET16(blitter_ram, A1_FPIXEL + 0);
+ int16_t a1_inc_x = (int16_t)GET16(blitter_ram, A1_INC + 2);
+ int16_t a1_inc_y = (int16_t)GET16(blitter_ram, A1_INC + 0);
+ uint16_t a1_incf_x = GET16(blitter_ram, A1_FINC + 2);
+ uint16_t a1_incf_y = GET16(blitter_ram, A1_FINC + 0);
+
+ int16_t a2_x = (int16_t)GET16(blitter_ram, A2_PIXEL + 2);
+ int16_t a2_y = (int16_t)GET16(blitter_ram, A2_PIXEL + 0);
+ uint16_t a2_mask_x = GET16(blitter_ram, A2_MASK + 2);
+ uint16_t a2_mask_y = GET16(blitter_ram, A2_MASK + 0);
+ int16_t a2_step_x = (int16_t)GET16(blitter_ram, A2_STEP + 2);
+ int16_t a2_step_y = (int16_t)GET16(blitter_ram, A2_STEP + 0);
+
+ uint64_t srcd1 = GET64(blitter_ram, SRCDATA);
+ uint64_t srcd2 = 0;
+ uint64_t dstd = GET64(blitter_ram, DSTDATA);
+ uint64_t patd = GET64(blitter_ram, PATTERNDATA);
+ uint32_t iinc = GET32(blitter_ram, INTENSITYINC);
+ uint64_t srcz1 = GET64(blitter_ram, SRCZINT);
+ uint64_t srcz2 = GET64(blitter_ram, SRCZFRAC);
+ uint64_t dstz = GET64(blitter_ram, DSTZ);
+ uint32_t zinc = GET32(blitter_ram, ZINC);
+ uint32_t collision = GET32(blitter_ram, COLLISIONCTRL);// 0=RESUME, 1=ABORT, 2=STOPEN
+
+ uint8_t pixsize = (dsta2 ? a2_pixsize : a1_pixsize); // From ACONTROL
+
+//Testing Trevor McFur--I *think* it's the circle on the lower RHS of the screen...
+/*logBlit = false;
+if (cmd == 0x05810601 && (GET16(blitter_ram, PIXLINECOUNTER + 2) == 96)
+ && (GET16(blitter_ram, PIXLINECOUNTER + 0) == 72))
+ logBlit = true;//*/
//Testing...
//if (cmd == 0x1401060C) patd = 0xFFFFFFFFFFFFFFFFLL;
//if (cmd == 0x1401060C) patd = 0x00000000000000FFLL;
So... It's reversed! The data organization of the patd register is [low 32][high 32]! !!! FIX !!! [DONE]
And fix all the other 64 bit registers [DONE]
*/
-
-if (cmd == 0x1401060C)
+/*if (cmd == 0x1401060C)
{
printf("logBlit = %s, cmd = %08X\n", (logBlit ? "T" : "F"), cmd);
fflush(stdout);
-}
+}*/
+/*logBlit = false;
+if ((cmd == 0x00010200) && (GET16(blitter_ram, PIXLINECOUNTER + 2) == 9))
+ logBlit = true;
+
+; Pink altimeter bar
+
+Blit! (00110000 <- 000BF010) count: 9 x 23, A1/2_FLAGS: 000042E2/00010020 [cmd: 00010200]
+ CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
+ A1 step values: -10 (X), 1 (Y)
+ A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 262/132, A2 x/y: 129/0
+;x-coord is 257 in pic, so add 5
+;20 for ship, 33 for #... Let's see if we can find 'em!
+
+; Black altimeter bar
+
+Blit! (00110000 <- 000BF010) count: 5 x 29, A1/2_FLAGS: 000042E2/00010020 [cmd: 00010200]
+ CMD -> src: dst: misc: a1ctl: UPDA1 mode: ity: PATDSEL z-op: op: LFU_CLEAR ctrl:
+ A1 step values: -8 (X), 1 (Y)
+ A1 -> pitch: 4 phrases, depth: 16bpp, z-off: 3, width: 320 (21), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 264/126, A2 x/y: 336/0
+
+Here's the pink bar--note that it's phrase mode without dread, so how does this work???
+Not sure, but I *think* that somehow it MUXes the data at the write site in on the left or right side
+of the write data when masked in phrase mode. I'll have to do some tracing to see if this is the mechanism
+it uses or not...
+
+Blit! (CMD = 00010200)
+Flags: UPDA1 PATDSEL
+ count = 9 x 11
+ a1_base = 00110010, a2_base = 000BD7E0
+ a1_x = 0106, a1_y = 0090, a1_frac_x = 0000, a1_frac_y = 8000, a2_x = 025A, a2_y = 0000
+ a1_step_x = FFF6, a1_step_y = 0001, a1_stepf_x = 5E00, a1_stepf_y = D100, a2_step_x = FFF7, a2_step_y = 0001
+ a1_inc_x = 0001, a1_inc_y = FFFF, a1_incf_x = 0000, a1_incf_y = E000
+ a1_win_x = 0000, a1_win_y = 0000, a2_mask_x = 0000, a2_mask_y = 0000
+ a2_mask=F a1add=+phr/+0 a2add=+1/+0
+ a1_pixsize = 4, a2_pixsize = 4
+ srcd=BAC673AC2C92E578 dstd=0000000000000000 patd=74C074C074C074C0 iinc=0002E398
+ srcz1=7E127E12000088DA srcz2=DBE06DF000000000 dstz=0000000000000000 zinc=FFFE4840, coll=0
+ Phrase mode is ON
+ [in=T a1f=F a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
+ Entering INNER state...
+ Entering DWRITE state...
+ Dest write address/pix address: 0016A830/0 [dstart=20 dend=40 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=0 daq=F] [7400000074C074C0] (icount=0007, inc=2)
+ Entering A1_ADD state [a1_x=0106, a1_y=0090, addasel=0, addbsel=0, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering DWRITE state...
+ Dest write address/pix address: 0016A850/0 [dstart=0 dend=40 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=0 daq=F] [74C074C074C074C0] (icount=0003, inc=4)
+ Entering A1_ADD state [a1_x=0108, a1_y=0090, addasel=0, addbsel=0, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering DWRITE state...
+ Dest write address/pix address: 0016A870/0 [dstart=0 dend=30 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=0 daq=F] [74C074C074C00000] (icount=FFFF, inc=4)
+ Entering A1_ADD state [a1_x=010C, a1_y=0090, addasel=0, addbsel=0, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering IDLE_INNER state...
+ Leaving INNER state... (ocount=000A)
+ [in=F a1f=F a1=T zf=F z=F a2=F iif=F iii=F izf=F izi=F]
+ Entering A1UPDATE state... (272/144 -> 262/145)
+ [in=T a1f=F a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
+ Entering INNER state...
+*/
+
// Bugs in Jaguar I
-
+
a2addy = a1addy; // A2 channel Y add bit is tied to A1's
//if (logBlit && (ocount > 20)) logBlit = false;
if (logBlit)
{
printf(" a1_base = %08X, a2_base = %08X\n", a1_base, a2_base);
-printf(" a1_x = %04X, a1_y = %04X, a1_frac_x = %04X, a1_frac_y = %04X, a2_x = %04X, a2_y = %04X\n", (uint16)a1_x, (uint16)a1_y, a1_frac_x, a1_frac_y, (uint16)a2_x, (uint16)a2_y);
-printf(" a1_step_x = %04X, a1_step_y = %04X, a1_stepf_x = %04X, a1_stepf_y = %04X, a2_step_x = %04X, a2_step_y = %04X\n", (uint16)a1_step_x, (uint16)a1_step_y, a1_stepf_x, a1_stepf_y, (uint16)a2_step_x, (uint16)a2_step_y);
-printf(" a1_inc_x = %04X, a1_inc_y = %04X, a1_incf_x = %04X, a1_incf_y = %04X\n", (uint16)a1_inc_x, (uint16)a1_inc_y, a1_incf_x, a1_incf_y);
+printf(" a1_x = %04X, a1_y = %04X, a1_frac_x = %04X, a1_frac_y = %04X, a2_x = %04X, a2_y = %04X\n", (uint16_t)a1_x, (uint16_t)a1_y, a1_frac_x, a1_frac_y, (uint16_t)a2_x, (uint16_t)a2_y);
+printf(" a1_step_x = %04X, a1_step_y = %04X, a1_stepf_x = %04X, a1_stepf_y = %04X, a2_step_x = %04X, a2_step_y = %04X\n", (uint16_t)a1_step_x, (uint16_t)a1_step_y, a1_stepf_x, a1_stepf_y, (uint16_t)a2_step_x, (uint16_t)a2_step_y);
+printf(" a1_inc_x = %04X, a1_inc_y = %04X, a1_incf_x = %04X, a1_incf_y = %04X\n", (uint16_t)a1_inc_x, (uint16_t)a1_inc_y, a1_incf_x, a1_incf_y);
printf(" a1_win_x = %04X, a1_win_y = %04X, a2_mask_x = %04X, a2_mask_y = %04X\n", a1_win_x, a1_win_y, a2_mask_x, a2_mask_y);
char x_add_str[4][4] = { "phr", "1", "0", "inc" };
printf(" a2_mask=%s a1add=%s%s/%s%s a2add=%s%s/%s%s\n", (a2_mask ? "T" : "F"), (a1xsign ? "-" : "+"), x_add_str[a1addx],
(a2ysign ? "-" : "+"), (a2addy ? "1" : "0"));
printf(" a1_pixsize = %u, a2_pixsize = %u\n", a1_pixsize, a2_pixsize);
printf(" srcd=%08X%08X dstd=%08X%08X patd=%08X%08X iinc=%08X\n",
- (uint32)(srcd1 >> 32), (uint32)(srcd1 & 0xFFFFFFFF),
- (uint32)(dstd >> 32), (uint32)(dstd & 0xFFFFFFFF),
- (uint32)(patd >> 32), (uint32)(patd & 0xFFFFFFFF), iinc);
+ (uint32_t)(srcd1 >> 32), (uint32_t)(srcd1 & 0xFFFFFFFF),
+ (uint32_t)(dstd >> 32), (uint32_t)(dstd & 0xFFFFFFFF),
+ (uint32_t)(patd >> 32), (uint32_t)(patd & 0xFFFFFFFF), iinc);
printf(" srcz1=%08X%08X srcz2=%08X%08X dstz=%08X%08X zinc=%08X, coll=%X\n",
- (uint32)(srcz1 >> 32), (uint32)(srcz1 & 0xFFFFFFFF),
- (uint32)(srcz2 >> 32), (uint32)(srcz2 & 0xFFFFFFFF),
- (uint32)(dstz >> 32), (uint32)(dstz & 0xFFFFFFFF), zinc, collision);
+ (uint32_t)(srcz1 >> 32), (uint32_t)(srcz1 & 0xFFFFFFFF),
+ (uint32_t)(srcz2 >> 32), (uint32_t)(srcz2 & 0xFFFFFFFF),
+ (uint32_t)(dstz >> 32), (uint32_t)(dstz & 0xFFFFFFFF), zinc, collision);
}
-#endif
+#endif
// Various state lines set up by user
fflush(stdout);
}
#endif
+//logBlit = false;
// Stopgap vars to simulate various lines
- uint16 a1FracCInX = 0, a1FracCInY = 0;
+ uint16_t a1FracCInX = 0, a1FracCInY = 0;
while (true)
{
// IDLE
-
+
if ((idle && !go) || (inner && outer0 && indone))
{
#ifdef VERBOSE_BLITTER_LOGGING
#endif
idlei = true;
- return;
+//Instead of a return, let's try breaking out of the loop...
+break;
+// return;
}
else
idlei = false;
-
+
// INNER LOOP ACTIVE
/*
Entering DWRITE state... (icount=0000, inc=4)
Now:
[in=F a1f=F a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
*/
-
+
if ((idle && go && !datinit)
|| (inner && !indone)
|| (inner && indone && !outer0 && !upda1f && !upda1 && notgzandp && !upda2 && !datinit)
}
else
inneri = false;
-
+
// A1 FRACTION UPDATE
-
+
if (inner && indone && !outer0 && upda1f)
{
a1fupdatei = true;
}
else
a2updatei = false;
-
+
// INITIALIZE INTENSITY FRACTION
-
+
if ((zupdate && !upda2 && datinit)
|| (a1update && !upda2 && datinit && notgzandp)
|| (inner && indone && !outer0 && !upda1f && !upda1 && notgzandp && !upda2 && datinit)
}
else
init_ifi = false;
-
+
// INITIALIZE INTENSITY INTEGER
-
+
if (init_if)
{
init_iii = true;
}
else
init_iii = false;
-
+
// INITIALIZE Z FRACTION
-
+
if (init_ii && gourz)
{
init_zfi = true;
}
else
init_zfi = false;
-
+
// INITIALIZE Z INTEGER
-
+
if (init_zf)
{
init_zii = true;
}
else
init_zii = false;
-
+
// Here we move the fooi into their foo counterparts in order to simulate the moving
// of data into the various FDSYNCs... Each time we loop we simulate one clock cycle...
fflush(stdout);
}
#endif
- uint16 icount = GET16(blitter_ram, PIXLINECOUNTER + 2);
+ uint16_t icount = GET16(blitter_ram, PIXLINECOUNTER + 2);
bool idle_inner = true, step = true, sreadx = false, szreadx = false, sread = false,
szread = false, dread = false, dzread = false, dwrite = false, dzwrite = false;
bool inner0 = false;
bool textext = false, txtread = false;
//other stuff
-uint8 srcshift = 0;
+uint8_t srcshift = 0;
bool sshftld = true; // D flipflop (D -> Q): instart -> sshftld
//NOTE: sshftld probably is only asserted at the beginning of the inner loop. !!! FIX !!!
/*
bool dsta_addi = (dwritei && !dstwrz) || dzwritei;
bool gensrc = sreadxi || szreadxi || sreadi || szreadi;
- bool gendst = dreadi || szreadi || dwritei || dzwritei;
+ bool gendst = dreadi || dzreadi || dwritei || dzwritei;
bool gena2i = (gensrc && !dsta2) || (gendst && dsta2);
bool zaddr = szreadx || szread || dzread || dzwrite;
The /a2update term on bits 0 and 1 is redundant.
Now look-ahead based
*/
- uint8 addasel = (a1fupdate || (a1_add && a1addx == 3) ? 0x01 : 0x00);
+ uint8_t addasel = (a1fupdate || (a1_add && a1addx == 3) ? 0x01 : 0x00);
addasel |= (a1_add && a1addx == 3 ? 0x02 : 0x00);
addasel |= (a2update ? 0x04 : 0x00);
/* Address adder input A X constant selection
similarly for A2
JLH: Also, 11 will likewise set the value to 111
*/
- uint8 a1_xconst = 6 - a1_pixsize, a2_xconst = 6 - a2_pixsize;
+ uint8_t a1_xconst = 6 - a1_pixsize, a2_xconst = 6 - a2_pixsize;
if (a1addx == 1)
a1_xconst = 0;
else if (a2addx & 0x02)
a2_xconst = 7;
- uint8 adda_xconst = (a2_add ? a2_xconst : a1_xconst);
+ uint8_t adda_xconst = (a2_add ? a2_xconst : a1_xconst);
/* Address adder input A Y constant selection
22 June 94 - This was erroneous, because only the a1addy bit was reflected here.
Therefore, the selection has to be controlled by a bug fix bit.
+ a1update . a1_stepld
+ a2update . a2_stepld
*/
- uint8 addbsel = (a2update || a2_add || (a1fupdate && a1_stepld)
+ uint8_t addbsel = (a2update || a2_add || (a1fupdate && a1_stepld)
|| (a1update && a1_stepld) || (a2update && a2_stepld) ? 0x01 : 0x00);
addbsel |= (a1fupdate || (a1_add && a1addx == 3) || (a1fupdate && a1_stepld)
|| (a1update && a1_stepld) || (a2update && a2_stepld) ? 0x02 : 0x00);
Masking is enabled for a1 when a1addx[0..1] is 00, and the value
is 6 - the pixel size (again!)
*/
- uint8 maska1 = (a1_add && a1addx == 0 ? 6 - a1_pixsize : 0);
- uint8 maska2 = (a2_add && a2addx == 0 ? 6 - a2_pixsize : 0);
- uint8 modx = (a2_add ? maska2 : maska1);
+ uint8_t maska1 = (a1_add && a1addx == 0 ? 6 - a1_pixsize : 0);
+ uint8_t maska2 = (a2_add && a2addx == 0 ? 6 - a2_pixsize : 0);
+ uint8_t modx = (a2_add ? maska2 : maska1);
/* Generate load strobes for the increment updates */
/*A1pldt := NAN2 (a1pldt, atick[1], a1_add);
Bit 2 = (gourd + gourz) . /(init_if + init_ii + init_zf + init_zi)
+ dwrite . srcshade
*/
-uint8 daddasel = ((dwrite && gourd) || (dzwrite && gourz) || istepadd || zstepfadd
+uint8_t daddasel = ((dwrite && gourd) || (dzwrite && gourz) || istepadd || zstepfadd
|| init_if || init_ii || init_zf || init_zi ? 0x01 : 0x00);
daddasel |= ((dzwrite && gourz) || zstepadd || zstepfadd ? 0x02 : 0x00);
daddasel |= (((gourd || gourz) && !(init_if || init_ii || init_zf || init_zi))
+ istepadd + istepfadd + zstepadd + zstepfadd
Bit 3 = istepadd + istepfadd + zstepadd + zstepfadd
*/
-uint8 daddbsel = ((dwrite && gourd) || (dzwrite && gourz) || (dwrite && srcshade)
+uint8_t daddbsel = ((dwrite && gourd) || (dzwrite && gourz) || (dwrite && srcshade)
|| istepadd || zstepadd || init_if || init_ii || init_zf || init_zi ? 0x01 : 0x00);
daddbsel |= ((dzwrite && gourz) || zstepadd || zstepfadd ? 0x02 : 0x00);
daddbsel |= ((dwrite && gourd) || (dzwrite && gourz) || (dwrite && srcshade)
/* Data adder mode control
000 16-bit normal add
001 16-bit saturating add with carry
-010 8-bit saturating add with carry, carry into top byte is
+010 8-bit saturating add with carry, carry into top byte is
inhibited (YCrCb)
-011 8-bit saturating add with carry, carry into top byte and
+011 8-bit saturating add with carry, carry into top byte and
between top nybbles is inhibited (CRY)
100 16-bit normal add with carry
101 16-bit saturating add
110 8-bit saturating add, carry into top byte is inhibited
-111 8-bit saturating add, carry into top byte and between top
+111 8-bit saturating add, carry into top byte and between top
nybbles is inhibited
The first five are used for Gouraud calculations, the latter three
+ init_ii . /topnen . /topben . /ext_int
+ init_ii . topnen . topben . /ext_int
+ init_zi
-
+
Bit 1 = dwrite . gourd . atick[1] . /topben . /ext_int
+ istepadd . /topben . /ext_int
+ /gourd . /gourz . /topben
+ istepadd . ext_int
+ init_ii . ext_int
*/
-uint8 daddmode = ((dzwrite && gourz) || (dwrite && gourd && !topnen && !topben && !ext_int)
+uint8_t daddmode = ((dzwrite && gourz) || (dwrite && gourd && !topnen && !topben && !ext_int)
|| (dwrite && gourd && topnen && topben && !ext_int) || zstepadd
|| (istepadd && !topnen && !topben && !ext_int)
|| (istepadd && topnen && topben && !ext_int) || (!gourd && !gourz && !topnen && !topben)
|| (init_ii && !topben && !ext_int) ? 0x02 : 0x00);
daddmode |= ((!gourd && !gourz) || shadeadd || (dwrite && gourd && ext_int)
|| (istepadd && ext_int) || (init_ii && ext_int) ? 0x04 : 0x00);
-/* Data add load controls
-Pattern fraction (dest data) is loaded on
+/* Data add load controls
+Pattern fraction (dest data) is loaded on
dwrite . gourd . atick[0]
+ istepfadd . /datinit
+ init_if
dwrite . gourd . atick[1]
+ istepadd . /datinit . /datinit
+ init_ii
-Source z1 is loaded on
+Source z1 is loaded on
dzwrite . gourz . atick[1]
+ zstepadd . /datinit . /datinit
+ init_zi
-Source z2 is loaded on
+Source z2 is loaded on
dzwrite . gourz . atick[0]
+ zstepfadd
+ init_zf
Bit 1 = adddsel
+ dzwrite1d
*/
-uint8 data_sel = ((!patdsel && !adddsel) || dzwrite ? 0x01 : 0x00)
+uint8_t data_sel = ((!patdsel && !adddsel) || dzwrite ? 0x01 : 0x00)
| (adddsel || dzwrite ? 0x02 : 0x00);
-uint32 address, pixAddr;
+uint32_t address, pixAddr;
ADDRGEN(address, pixAddr, gena2i, zaddr,
a1_x, a1_y, a1_base, a1_pitch, a1_pixsize, a1_width, a1_zoffset,
a2_x, a2_y, a2_base, a2_pitch, a2_pixsize, a2_width, a2_zoffset);
/* Generate source alignment shift
-------------------------------
-The source alignment shift for data move is the difference between
-the source and destination X pointers, multiplied by the pixel
-size. Only the low six bits of the pointers are of interest, as
-pixel sizes are always a power of 2 and window rows are always
-phrase aligned.
+The source alignment shift for data move is the difference between
+the source and destination X pointers, multiplied by the pixel
+size. Only the low six bits of the pointers are of interest, as
+pixel sizes are always a power of 2 and window rows are always
+phrase aligned.
When not in phrase mode, the top 3 bits of the shift value are
set to zero (2/26).
Source shifting is also used to extract bits for bit-to-byte
-expansion in phrase mode. This involves only the bottom three
+expansion in phrase mode. This involves only the bottom three
bits of the shift value, and is based on the offset within the
phrase of the destination X pointer, in pixels.
Source shifting is disabled when srcen is not set.
*/
-uint8 dstxp = (dsta2 ? a2_x : a1_x) & 0x3F;
-uint8 srcxp = (dsta2 ? a1_x : a2_x) & 0x3F;
-uint8 shftv = ((dstxp - srcxp) << pixsize) & 0x3F;
+uint8_t dstxp = (dsta2 ? a2_x : a1_x) & 0x3F;
+uint8_t srcxp = (dsta2 ? a1_x : a2_x) & 0x3F;
+uint8_t shftv = ((dstxp - srcxp) << pixsize) & 0x3F;
/* The phrase mode alignment count is given by the phrase offset
of the first pixel, for bit to byte expansion */
-uint8 pobb = 0;
+uint8_t pobb = 0;
if (pixsize == 3)
pobb = dstxp & 0x07;
pobb = dstxp & 0x01;
bool pobbsel = phrase_mode && bcompen;
-uint8 loshd = (pobbsel ? pobb : shftv) & 0x07;
-uint8 shfti = (srcen || pobbsel ? (sshftld ? loshd : srcshift & 0x07) : 0);
+uint8_t loshd = (pobbsel ? pobb : shftv) & 0x07;
+uint8_t shfti = (srcen || pobbsel ? (sshftld ? loshd : srcshift & 0x07) : 0);
/* Enable for high bits is srcen . phrase_mode */
shfti |= (srcen && phrase_mode ? (sshftld ? shftv & 0x38 : srcshift & 0x38) : 0);
srcshift = shfti;
fflush(stdout);
}
#endif
-//uint32 srcAddr, pixAddr;
+//uint32_t srcAddr, pixAddr;
//ADDRGEN(srcAddr, pixAddr, gena2i, zaddr,
// a1_x, a1_y, a1_base, a1_pitch, a1_pixsize, a1_width, a1_zoffset,
// a2_x, a2_y, a2_base, a2_pitch, a2_pixsize, a2_width, a2_zoffset);
-srcd2 = srcd1;
-srcd1 = ((uint64)JaguarReadLong(address, BLITTER) << 32) | (uint64)JaguarReadLong(address + 4, BLITTER);
+ srcd2 = srcd1;
+ srcd1 = ((uint64_t)JaguarReadLong(address + 0, BLITTER) << 32)
+ | (uint64_t)JaguarReadLong(address + 4, BLITTER);
//Kludge to take pixel size into account...
+//Hmm. If we're not in phrase mode, this is most likely NOT going to be used...
+//Actually, it would be--because of BCOMPEN expansion, for example...
if (!phrase_mode)
{
- if (pixsize == 5)
- srcd1 >>= 32;
- else if (pixsize == 4)
- srcd1 >>= 48;
- else
+ if (bcompen)
srcd1 >>= 56;
-}
+ else
+ {
+ if (pixsize == 5)
+ srcd1 >>= 32;
+ else if (pixsize == 4)
+ srcd1 >>= 48;
+ else
+ srcd1 >>= 56;
+ }
+}//*/
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
printf(" Source extra read address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
- (uint32)(srcd1 >> 32), (uint32)(srcd1 & 0xFFFFFFFF));
+ (uint32_t)(srcd1 >> 32), (uint32_t)(srcd1 & 0xFFFFFFFF));
fflush(stdout);
}
#endif
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" Entering SZREADX state...\n");
+printf(" Entering SZREADX state...");
fflush(stdout);
}
+#endif
+ srcz2 = srcz1;
+ srcz1 = ((uint64_t)JaguarReadLong(address, BLITTER) << 32) | (uint64_t)JaguarReadLong(address + 4, BLITTER);
+#ifdef VERBOSE_BLITTER_LOGGING
+if (logBlit)
+{
+ printf(" Src Z extra read address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
+ (uint32_t)(dstz >> 32), (uint32_t)(dstz & 0xFFFFFFFF));
+ fflush(stdout);
+}
#endif
}
fflush(stdout);
}
#endif
-//uint32 srcAddr, pixAddr;
+//uint32_t srcAddr, pixAddr;
//ADDRGEN(srcAddr, pixAddr, gena2i, zaddr,
// a1_x, a1_y, a1_base, a1_pitch, a1_pixsize, a1_width, a1_zoffset,
// a2_x, a2_y, a2_base, a2_pitch, a2_pixsize, a2_width, a2_zoffset);
srcd2 = srcd1;
-srcd1 = ((uint64)JaguarReadLong(address, BLITTER) << 32) | (uint64)JaguarReadLong(address + 4, BLITTER);
+srcd1 = ((uint64_t)JaguarReadLong(address, BLITTER) << 32) | (uint64_t)JaguarReadLong(address + 4, BLITTER);
//Kludge to take pixel size into account...
if (!phrase_mode)
{
- if (pixsize == 5)
- srcd1 >>= 32;
- else if (pixsize == 4)
- srcd1 >>= 48;
- else
+ if (bcompen)
srcd1 >>= 56;
+ else
+ {
+ if (pixsize == 5)
+ srcd1 >>= 32;
+ else if (pixsize == 4)
+ srcd1 >>= 48;
+ else
+ srcd1 >>= 56;
+ }
}
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" Source read address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
- (uint32)(srcd1 >> 32), (uint32)(srcd1 & 0xFFFFFFFF));
+printf(" Source read address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
+ (uint32_t)(srcd1 >> 32), (uint32_t)(srcd1 & 0xFFFFFFFF));
fflush(stdout);
}
#endif
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" Entering SZREADX state...\n");
+printf(" Entering SZREAD state...");
fflush(stdout);
}
+#endif
+ srcz2 = srcz1;
+ srcz1 = ((uint64_t)JaguarReadLong(address, BLITTER) << 32) | (uint64_t)JaguarReadLong(address + 4, BLITTER);
+//Kludge to take pixel size into account... I believe that it only has to take 16BPP mode into account. Not sure tho.
+if (!phrase_mode && pixsize == 4)
+ srcz1 >>= 48;
+
+#ifdef VERBOSE_BLITTER_LOGGING
+if (logBlit)
+{
+ printf(" Src Z read address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
+ (uint32_t)(dstz >> 32), (uint32_t)(dstz & 0xFFFFFFFF));
+ fflush(stdout);
+}
#endif
}
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" Entering DREAD state...\n");
+printf(" Entering DREAD state...");
fflush(stdout);
}
#endif
-//uint32 dstAddr, pixAddr;
+//uint32_t dstAddr, pixAddr;
//ADDRGEN(dstAddr, pixAddr, gena2i, zaddr,
// a1_x, a1_y, a1_base, a1_pitch, a1_pixsize, a1_width, a1_zoffset,
// a2_x, a2_y, a2_base, a2_pitch, a2_pixsize, a2_width, a2_zoffset);
-dstd = ((uint64)JaguarReadLong(address, BLITTER) << 32) | (uint64)JaguarReadLong(address + 4, BLITTER);
+dstd = ((uint64_t)JaguarReadLong(address, BLITTER) << 32) | (uint64_t)JaguarReadLong(address + 4, BLITTER);
//Kludge to take pixel size into account...
if (!phrase_mode)
{
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" Dest read address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
- (uint32)(dstd >> 32), (uint32)(dstd & 0xFFFFFFFF));
+printf(" Dest read address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
+ (uint32_t)(dstd >> 32), (uint32_t)(dstd & 0xFFFFFFFF));
fflush(stdout);
}
#endif
fflush(stdout);
}
#endif
- dstz = ((uint64)JaguarReadLong(address, BLITTER) << 32) | (uint64)JaguarReadLong(address + 4, BLITTER);
+ dstz = ((uint64_t)JaguarReadLong(address, BLITTER) << 32) | (uint64_t)JaguarReadLong(address + 4, BLITTER);
+//Kludge to take pixel size into account... I believe that it only has to take 16BPP mode into account. Not sure tho.
+if (!phrase_mode && pixsize == 4)
+ dstz >>= 48;
+
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
printf(" Dest Z read address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
- (uint32)(dstz >> 32), (uint32)(dstz & 0xFFFFFFFF));
+ (uint32_t)(dstz >> 32), (uint32_t)(dstz & 0xFFFFFFFF));
fflush(stdout);
}
#endif
}
+// These vars should probably go further up in the code... !!! FIX !!!
+// We can't preassign these unless they're static...
+//uint64_t srcz = 0; // These are assigned to shut up stupid compiler warnings--dwrite is ALWAYS asserted
+//bool winhibit = false;
+uint64_t srcz;
+bool winhibit;
//NOTE: SRCSHADE requires GOURZ to be set to work properly--another Jaguar I bug
if (dwrite)
{
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" Entering DWRITE state...\n");
+printf(" Entering DWRITE state...");
fflush(stdout);
}
#endif
//Counter is done on the dwrite state...! (We'll do it first, since it affects dstart/dend calculations.)
//Here's the voodoo for figuring the correct amount of pixels in phrase mode (or not):
- int8 inct = -((dsta2 ? a2_x : a1_x) & 0x07); // From INNER_CNT
- uint8 inc = 0;
+ int8_t inct = -((dsta2 ? a2_x : a1_x) & 0x07); // From INNER_CNT
+ uint8_t inc = 0;
inc = (!phrase_mode || (phrase_mode && (inct & 0x01)) ? 0x01 : 0x00);
inc |= (phrase_mode && (((pixsize == 3 || pixsize == 4) && (inct & 0x02)) || pixsize == 5 && !(inct & 0x01)) ? 0x02 : 0x00);
inc |= (phrase_mode && ((pixsize == 3 && (inct & 0x04)) || (pixsize == 4 && !(inct & 0x03))) ? 0x04 : 0x00);
inc |= (phrase_mode && pixsize == 3 && !(inct & 0x07) ? 0x08 : 0x00);
- uint16 oldicount = icount; // Save icount to detect underflow...
+ uint16_t oldicount = icount; // Save icount to detect underflow...
icount -= inc;
if (icount == 0 || ((icount & 0x8000) && !(oldicount & 0x8000)))
//Start & end write mask computations...
//*********************************************************************************
-uint8 dstart = 0;
+uint8_t dstart = 0;
if (pixsize == 3)
dstart = (dstxp & 0x07) << 3;
dstart = (phrase_mode ? dstart : pixAddr & 0x07);
//This is the other Jaguar I bug... Normally, should ALWAYS select a1_x here.
-uint16 dstxwr = (dsta2 ? a2_x : a1_x) & 0x7FFE;
-uint16 pseq = dstxwr ^ (a1_win_x & 0x7FFE);
+uint16_t dstxwr = (dsta2 ? a2_x : a1_x) & 0x7FFE;
+uint16_t pseq = dstxwr ^ (a1_win_x & 0x7FFE);
pseq = (pixsize == 5 ? pseq : pseq & 0x7FFC);
pseq = ((pixsize & 0x06) == 4 ? pseq : pseq & 0x7FF8);
bool penden = clip_a1 && (pseq == 0);
-uint8 window_mask = 0;
+uint8_t window_mask = 0;
if (pixsize == 3)
window_mask = (a1_win_x & 0x07) << 3;
/* The mask to be used if within one phrase of the end of the inner
loop, similarly */
-uint8 inner_mask = 0;
+uint8_t inner_mask = 0;
if (pixsize == 3)
inner_mask = (icount & 0x07) << 3;
the inner mask, where is all cases 000 means 1000. */
window_mask = (window_mask == 0 ? 0x40 : window_mask);
inner_mask = (inner_mask == 0 ? 0x40 : inner_mask);
-uint8 emask = (window_mask > inner_mask ? inner_mask : window_mask);
+uint8_t emask = (window_mask > inner_mask ? inner_mask : window_mask);
/* The mask to be used for the pixel size, to which must be added
the bit offset */
-uint8 pma = pixAddr + (1 << pixsize);
+uint8_t pma = pixAddr + (1 << pixsize);
/* Select the mask */
-uint8 dend = (phrase_mode ? emask : pma);
+uint8_t dend = (phrase_mode ? emask : pma);
/* The cycle width in phrase mode is normally one phrase. However,
at the start and end it may be narrower. The start and end masks
This is only used for writes in phrase mode.
Start and end from the address level of the pipeline are used.
*/
-uint8 pwidth = (((dend | dstart) & 0x07) == 0 ? 0x08 : (dend - dstart) & 0x07);
+uint8_t pwidth = (((dend | dstart) & 0x07) == 0 ? 0x08 : (dend - dstart) & 0x07);
-//uint32 dstAddr, pixAddr;
+//uint32_t dstAddr, pixAddr;
//ADDRGEN(dstAddr, pixAddr, gena2i, zaddr,
// a1_x, a1_y, a1_base, a1_pitch, a1_pixsize, a1_width, a1_zoffset,
// a2_x, a2_y, a2_base, a2_pitch, a2_pixsize, a2_width, a2_zoffset);
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" Dest write address/pix address: %08X/%1X", address, pixAddr);
-printf(" [dstart=%X dend=%X pwidth=%X srcshift=%X]", dstart, dend, pwidth, srcshift);
-fflush(stdout);
-printf("[daas=%X dabs=%X dam=%X ds=%X daq=%s]", daddasel, daddbsel, daddmode, data_sel, (daddq_sel ? "T" : "F"));
-fflush(stdout);
+ printf(" Dest write address/pix address: %08X/%1X", address, pixAddr);
+ fflush(stdout);
}
#endif
+//More testing... This is almost certainly wrong, but how else does this work???
+//Seems to kinda work... But still, this doesn't seem to make any sense!
+if (phrase_mode && !dsten)
+ dstd = ((uint64_t)JaguarReadLong(address, BLITTER) << 32) | (uint64_t)JaguarReadLong(address + 4, BLITTER);
+
//Testing only... for now...
//This is wrong because the write data is a combination of srcd and dstd--either run
-//thru the LFU or in PATDSEL or ADDDSEL mode.
+//thru the LFU or in PATDSEL or ADDDSEL mode. [DONE now, thru DATA module]
// Precedence is ADDDSEL > PATDSEL > LFU.
//Also, doesn't take into account the start & end masks, or the phrase width...
//Now it does!
// srcd2 = xxxx xxxx 0123 4567, srcd = 8901 2345 xxxx xxxx, srcshift = $20 (32)
-uint64 srcd = (srcd2 << (64 - srcshift)) | (srcd1 >> srcshift);
+uint64_t srcd = (srcd2 << (64 - srcshift)) | (srcd1 >> srcshift);
//bleh, ugly ugly ugly
if (srcshift == 0)
srcd = srcd1;
-//Temporary kludge, to see if the fractional pattern does anything...
-//This works, BTW
-if (patfadd)
+//NOTE: This only works with pixel sizes less than 8BPP...
+//DOUBLE NOTE: Still need to do regression testing to ensure that this doesn't break other stuff... !!! CHECK !!!
+if (!phrase_mode && srcshift != 0)
+ srcd = ((srcd2 & 0xFF) << (8 - srcshift)) | ((srcd1 & 0xFF) >> srcshift);
+
+//Z DATA() stuff done here... And it has to be done before any Z shifting...
+//Note that we need to have phrase mode start/end support here... (Not since we moved it from dzwrite...!)
+/*
+Here are a couple of Cybermorph blits with Z:
+$00113078 // DSTEN DSTENZ DSTWRZ CLIP_A1 GOURD GOURZ PATDSEL ZMODE=4
+$09900F39 // SRCEN DSTEN DSTENZ DSTWRZ UPDA1 UPDA1F UPDA2 DSTA2 ZMODE=4 LFUFUNC=C DCOMPEN
+
+We're having the same phrase mode overwrite problem we had with the pixels... !!! FIX !!!
+Odd. It's equating 0 with 0... Even though ZMODE is $04 (less than)!
+*/
+if (gourz)
{
- uint16 addq[4];
- uint8 initcin[4] = { 0, 0, 0, 0 };
- ADDARRAY(addq, 4/*daddasel*/, 4/*daddbsel*/, 0/*daddmode*/, dstd, iinc, initcin, 0, 0, 0, patd, srcd, 0, 0, 0, 0);
- srcd1 = ((uint64)addq[3] << 48) | ((uint64)addq[2] << 32) | ((uint64)addq[1] << 16) | (uint64)addq[0];
+/*
+void ADDARRAY(uint16_t * addq, uint8_t daddasel, uint8_t daddbsel, uint8_t daddmode,
+ uint64_t dstd, uint32_t iinc, uint8_t initcin[], uint64_t initinc, uint16_t initpix,
+ uint32_t istep, uint64_t patd, uint64_t srcd, uint64_t srcz1, uint64_t srcz2,
+ uint32_t zinc, uint32_t zstep)
+*/
+ uint16_t addq[4];
+ uint8_t initcin[4] = { 0, 0, 0, 0 };
+ ADDARRAY(addq, 7/*daddasel*/, 6/*daddbsel*/, 0/*daddmode*/, 0, 0, initcin, 0, 0, 0, 0, 0, srcz1, srcz2, zinc, 0);
+ srcz2 = ((uint64_t)addq[3] << 48) | ((uint64_t)addq[2] << 32) | ((uint64_t)addq[1] << 16) | (uint64_t)addq[0];
+ ADDARRAY(addq, 6/*daddasel*/, 7/*daddbsel*/, 1/*daddmode*/, 0, 0, initcin, 0, 0, 0, 0, 0, srcz1, srcz2, zinc, 0);
+ srcz1 = ((uint64_t)addq[3] << 48) | ((uint64_t)addq[2] << 32) | ((uint64_t)addq[1] << 16) | (uint64_t)addq[0];
+
+#if 0//def VERBOSE_BLITTER_LOGGING
+if (logBlit)
+{
+ printf("\n[srcz1=%08X%08X, srcz2=%08X%08X, zinc=%08X",
+ (uint32_t)(srcz1 >> 32), (uint32_t)(srcz1 & 0xFFFFFFFF),
+ (uint32_t)(srcz2 >> 32), (uint32_t)(srcz2 & 0xFFFFFFFF), zinc);
+ fflush(stdout);
+}
+#endif
}
-//Note that we still don't take atick[0] & [1] into account here, so this will skip half of the data needed... !!! FIX !!!
-//Not yet enumerated: dbinh, srcdread, srczread
+uint8_t zSrcShift = srcshift & 0x30;
+srcz = (srcz2 << (64 - zSrcShift)) | (srcz1 >> zSrcShift);
+//bleh, ugly ugly ugly
+if (zSrcShift == 0)
+ srcz = srcz1;
+
+#if 0//def VERBOSE_BLITTER_LOGGING
+if (logBlit)
+{
+ printf(" srcz=%08X%08X]\n", (uint32_t)(srcz >> 32), (uint32_t)(srcz & 0xFFFFFFFF));
+ fflush(stdout);
+}
+#endif
//When in SRCSHADE mode, it adds the IINC to the read source (from LFU???)
//According to following line, it gets LFU mode. But does it feed the source into the LFU
//NOTE: This is basically doubling the work done by DATA--since this is what
// ADDARRAY is loaded with when srschshade is enabled... !!! FIX !!!
// Also note that it doesn't work properly unless GOURZ is set--there's the clue!
- uint16 addq[4];
- uint8 initcin[4] = { 0, 0, 0, 0 };
+ uint16_t addq[4];
+ uint8_t initcin[4] = { 0, 0, 0, 0 };
ADDARRAY(addq, 4/*daddasel*/, 5/*daddbsel*/, 7/*daddmode*/, dstd, iinc, initcin, 0, 0, 0, patd, srcd, 0, 0, 0, 0);
- srcd = ((uint64)addq[3] << 48) | ((uint64)addq[2] << 32) | ((uint64)addq[1] << 16) | (uint64)addq[0];
+ srcd = ((uint64_t)addq[3] << 48) | ((uint64_t)addq[2] << 32) | ((uint64_t)addq[1] << 16) | (uint64_t)addq[0];
}
//Seems to work... Not 100% sure tho.
//end try this
-bool winhibit;// = false;
-uint64 wdata;
-uint8 dcomp, zcomp;
+//Temporary kludge, to see if the fractional pattern does anything...
+//This works, BTW
+//But it seems to mess up in Cybermorph... the shading should be smooth but it isn't...
+//Seems the carry out is lost again... !!! FIX !!! [DONE--see below]
+if (patfadd)
+{
+ uint16_t addq[4];
+ uint8_t initcin[4] = { 0, 0, 0, 0 };
+ ADDARRAY(addq, 4/*daddasel*/, 4/*daddbsel*/, 0/*daddmode*/, dstd, iinc, initcin, 0, 0, 0, patd, srcd, 0, 0, 0, 0);
+ srcd1 = ((uint64_t)addq[3] << 48) | ((uint64_t)addq[2] << 32) | ((uint64_t)addq[1] << 16) | (uint64_t)addq[0];
+}
+
+//Note that we still don't take atick[0] & [1] into account here, so this will skip half of the data needed... !!! FIX !!!
+//Not yet enumerated: dbinh, srcdread, srczread
+//Also, should do srcshift on the z value in phrase mode... !!! FIX !!! [DONE]
+//As well as add a srcz variable we can set external to this state... !!! FIX !!! [DONE]
+
+uint64_t wdata;
+uint8_t dcomp, zcomp;
DATA(wdata, dcomp, zcomp, winhibit,
true, cmpdst, daddasel, daddbsel, daddmode, daddq_sel, data_sel, 0/*dbinh*/,
dend, dstart, dstd, iinc, lfufunc, patd, patdadd,
phrase_mode, srcd, false/*srcdread*/, false/*srczread*/, srcz2add, zmode,
- bcompen, bkgwren, dcompen, icount & 0x07, pixsize);
-// bool bcompen, bool bkgwren, bool dcompen, uint8 icount, uint8 pixsize)
+ bcompen, bkgwren, dcompen, icount & 0x07, pixsize,
+ srcz, dstz, zinc);
+/*
+Seems that the phrase mode writes with DCOMPEN and DSTEN are corrupting inside of DATA: !!! FIX !!!
+It's fairly random as well. 7CFE -> 7DFE, 7FCA -> 78CA, 7FA4 -> 78A4, 7F88 -> 8F88
+It could be related to an uninitialized variable, like the zmode bug...
+[DONE]
+It was a bug in the dech38el data--it returned $FF for ungated instead of $00...
+
+Blit! (CMD = 09800609)
+Flags: SRCEN DSTEN UPDA1 UPDA2 LFUFUNC=C DCOMPEN
+ count = 10 x 12
+ a1_base = 00110000, a2_base = 0010B2A8
+ a1_x = 004B, a1_y = 00D8, a1_frac_x = 0000, a1_frac_y = 0000, a2_x = 0704, a2_y = 0000
+ a1_step_x = FFF3, a1_step_y = 0001, a1_stepf_x = 0000, a1_stepf_y = 0000, a2_step_x = FFFC, a2_step_y = 0000
+ a1_inc_x = 0000, a1_inc_y = 0000, a1_incf_x = 0000, a1_incf_y = 0000
+ a1_win_x = 0000, a1_win_y = 0000, a2_mask_x = 0000, a2_mask_y = 0000
+ a2_mask=F a1add=+phr/+0 a2add=+phr/+0
+ a1_pixsize = 4, a2_pixsize = 4
+ srcd=0000000000000000 dstd=0000000000000000 patd=0000000000000000 iinc=00000000
+ srcz1=0000000000000000 srcz2=0000000000000000 dstz=0000000000000000 zinc=00000000, coll=0
+ Phrase mode is ON
+ [in=T a1f=F a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
+ Entering INNER state...
+ Entering SREAD state... Source read address/pix address: 0010C0B0/0 [0000000078047804]
+ Entering A2_ADD state [a2_x=0704, a2_y=0000, addasel=0, addbsel=1, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering DREAD state...
+ Dest read address/pix address: 00197240/0 [0000000000000028]
+ Entering DWRITE state...
+ Dest write address/pix address: 00197240/0 [dstart=30 dend=40 pwidth=8 srcshift=30][daas=0 dabs=0 dam=7 ds=1 daq=F] [0000000000000028] (icount=0009, inc=1)
+ Entering A1_ADD state [a1_x=004B, a1_y=00D8, addasel=0, addbsel=0, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering SREAD state... Source read address/pix address: 0010C0B8/0 [7804780478047804]
+ Entering A2_ADD state [a2_x=0708, a2_y=0000, addasel=0, addbsel=1, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering DREAD state...
+ Dest read address/pix address: 00197260/0 [0028000000200008]
+ Entering DWRITE state...
+ Dest write address/pix address: 00197260/0 [dstart=0 dend=40 pwidth=8 srcshift=30][daas=0 dabs=0 dam=7 ds=1 daq=F] [0028780478047804] (icount=0005, inc=4)
+ Entering A1_ADD state [a1_x=004C, a1_y=00D8, addasel=0, addbsel=0, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering SREAD state... Source read address/pix address: 0010C0C0/0 [0000000000000000]
+ Entering A2_ADD state [a2_x=070C, a2_y=0000, addasel=0, addbsel=1, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering DREAD state...
+ Dest read address/pix address: 00197280/0 [0008001800180018]
+ Entering DWRITE state...
+ Dest write address/pix address: 00197280/0 [dstart=0 dend=40 pwidth=8 srcshift=30][daas=0 dabs=0 dam=7 ds=1 daq=F] [7804780478040018] (icount=0001, inc=4)
+ Entering A1_ADD state [a1_x=0050, a1_y=00D8, addasel=0, addbsel=0, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering SREAD state... Source read address/pix address: 0010C0C8/0 [000078047BFE7BFE]
+ Entering A2_ADD state [a2_x=0710, a2_y=0000, addasel=0, addbsel=1, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering DREAD state...
+ Dest read address/pix address: 001972A0/0 [0008002000000000]
+ Entering DWRITE state...
+ Dest write address/pix address: 001972A0/0 [dstart=0 dend=10 pwidth=8 srcshift=30][daas=0 dabs=0 dam=7 ds=1 daq=F] [0008002000000000] (icount=FFFD, inc=4)
+ Entering A1_ADD state [a1_x=0054, a1_y=00D8, addasel=0, addbsel=0, modx=2, addareg=F, adda_xconst=2, adda_yconst=0]...
+ Entering IDLE_INNER state...
+*/
-if (patdadd)
- patd = wdata;
+//Why isn't this taken care of in DATA? Because, DATA is modifying its local copy instead of the one used here.
+//!!! FIX !!! [DONE]
+//if (patdadd)
+// patd = wdata;
//if (patfadd)
// srcd1 = wdata;
A1_outside := OR6 (a1_outside, a1_x{15}, a1xgr, a1xeq, a1_y{15}, a1ygr, a1yeq);
*/
//NOTE: There seems to be an off-by-one bug here in the clip_a1 section... !!! FIX !!!
+// Actually, seems to be related to phrase mode writes...
+// Or is it? Could be related to non-15-bit compares as above?
if (clip_a1 && ((a1_x & 0x8000) || (a1_y & 0x8000) || (a1_x >= a1_win_x) || (a1_y >= a1_win_y)))
winhibit = true;
-/*if (dcompen)
-{
-//This is currently not correct for phrase mode. !!! FIX !!!
- if ((pixsize == 3 && (dcomp & 0x01))
- || (pixsize == 4 && (dcomp & 0x03)))
- winhibit = true;
-}*/
-
if (!winhibit)
{
if (phrase_mode)
else
JaguarWriteByte(address, wdata & 0x000000FF, BLITTER);
}
+}
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
- printf(" [%08X%08X]", (uint32)(wdata >> 32), (uint32)(wdata & 0xFFFFFFFF));
+ printf(" [%08X%08X]%s", (uint32_t)(wdata >> 32), (uint32_t)(wdata & 0xFFFFFFFF), (winhibit ? "[X]" : ""));
+ printf(" (icount=%04X, inc=%u)\n", icount, (uint16_t)inc);
+ printf(" [dstart=%X dend=%X pwidth=%X srcshift=%X]", dstart, dend, pwidth, srcshift);
+ printf("[daas=%X dabs=%X dam=%X ds=%X daq=%s]\n", daddasel, daddbsel, daddmode, data_sel, (daddq_sel ? "T" : "F"));
fflush(stdout);
}
#endif
-}
+ }
+ if (dzwrite)
+ {
+// OK, here's the big insight: When NOT in GOURZ mode, srcz1 & 2 function EXACTLY the same way that
+// srcd1 & 2 work--there's an implicit shift from srcz1 to srcz2 whenever srcz1 is read.
+// OTHERWISE, srcz1 is the integer for the computed Z and srcz2 is the fractional part.
+// Writes to srcz1 & 2 follow the same pattern as the other 64-bit registers--low 32 at the low address,
+// high 32 at the high address (little endian!).
+// NOTE: GOURZ is still not properly supported. Check patd/patf handling...
+// Phrase mode start/end masks are not properly supported either...
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" (icount=%04X, inc=%u)\n", icount, (uint16)inc);
-fflush(stdout);
+ printf(" Entering DZWRITE state...");
+ printf(" Dest Z write address/pix address: %08X/%1X [%08X%08X]\n", address, pixAddr,
+ (uint32_t)(srcz >> 32), (uint32_t)(srcz & 0xFFFFFFFF));
+ fflush(stdout);
}
#endif
- }
-
- if (dzwrite)
- {
+//This is not correct... !!! FIX !!!
+//Should be OK now... We'll see...
+//Nope. Having the same starstep write problems in phrase mode as we had with pixels... !!! FIX !!!
+//This is not causing the problem in Hover Strike... :-/
+//The problem was with the SREADX not shifting. Still problems with Z comparisons & other text in pregame screen...
+if (!winhibit)
+{
+ if (phrase_mode)
+ {
+ JaguarWriteLong(address + 0, srcz >> 32, BLITTER);
+ JaguarWriteLong(address + 4, srcz & 0xFFFFFFFF, BLITTER);
+ }
+ else
+ {
+ if (pixsize == 4)
+ JaguarWriteWord(address, srcz & 0x0000FFFF, BLITTER);
+ }
+}//*/
#ifdef VERBOSE_BLITTER_LOGGING
if (logBlit)
{
-printf(" Entering DZWRITE state...\n");
-fflush(stdout);
+// printf(" [%08X%08X]\n", (uint32_t)(srcz >> 32), (uint32_t)(srcz & 0xFFFFFFFF));
+// fflush(stdout);
+//printf(" [dstart=%X dend=%X pwidth=%X srcshift=%X]", dstart, dend, pwidth, srcshift);
+ printf(" [dstart=? dend=? pwidth=? srcshift=%X]", srcshift);
+ printf("[daas=%X dabs=%X dam=%X ds=%X daq=%s]\n", daddasel, daddbsel, daddmode, data_sel, (daddq_sel ? "T" : "F"));
+ fflush(stdout);
}
#endif
}
+/*
+This is because the address generator was using only 15 bits of the X when it should have
+used 16!
+
+There's a slight problem here: The X pointer isn't wrapping like it should when it hits
+the edge of the window... Notice how the X isn't reset at the edge of the window:
+
+Blit! (CMD = 00010000)
+Flags: PATDSEL
+ count = 160 x 261
+ a1_base = 000E8008, a2_base = 0001FA68
+ a1_x = 0000, a1_y = 0000, a1_frac_x = 0000, a1_frac_y = 0000, a2_x = 0000, a2_y = 0000
+ a1_step_x = 0000, a1_step_y = 0000, a1_stepf_x = 0000, a1_stepf_y = 0000, a2_step_x = 0000, a2_step_y = 0000
+ a1_inc_x = 0000, a1_inc_y = 0000, a1_incf_x = 0000, a1_incf_y = 0000
+ a1_win_x = 0000, a1_win_y = 0000, a2_mask_x = 0000, a2_mask_y = 0000
+ a2_mask=F a1add=+phr/+0 a2add=+phr/+0
+ a1_pixsize = 5, a2_pixsize = 5
+ srcd=7717771777177717 dstd=0000000000000000 patd=7730773077307730 iinc=00000000
+ srcz1=0000000000000000 srcz2=0000000000000000 dstz=0000000000000000 zinc=00000000, coll=0
+ Phrase mode is ON
+ [in=T a1f=F a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
+ Entering INNER state...
+ Entering DWRITE state... Dest write address/pix address: 000E8008/0 [7730773077307730] (icount=009E, inc=2)
+ srcz=0000000000000000][dcomp=AA zcomp=00 dbinh=00]
+[srcz=0000000000000000 dstz=0000000000000000 zwdata=0000000000000000 mask=7FFF]
+ [dstart=0 dend=40 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=0 daq=F]
+ Entering A1_ADD state [a1_x=0000, a1_y=0000, addasel=0, addbsel=0, modx=1, addareg=F, adda_xconst=1, adda_yconst=0]...
+ Entering DWRITE state... Dest write address/pix address: 000E8018/0 [7730773077307730] (icount=009C, inc=2)
+ srcz=0000000000000000][dcomp=AA zcomp=00 dbinh=00]
+[srcz=0000000000000000 dstz=0000000000000000 zwdata=0000000000000000 mask=7FFF]
+ [dstart=0 dend=40 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=0 daq=F]
+ Entering A1_ADD state [a1_x=0002, a1_y=0000, addasel=0, addbsel=0, modx=1, addareg=F, adda_xconst=1, adda_yconst=0]...
+
+...
+
+ Entering A1_ADD state [a1_x=009C, a1_y=0000, addasel=0, addbsel=0, modx=1, addareg=F, adda_xconst=1, adda_yconst=0]...
+ Entering DWRITE state... Dest write address/pix address: 000E84F8/0 [7730773077307730] (icount=0000, inc=2)
+ srcz=0000000000000000][dcomp=AA zcomp=00 dbinh=00]
+[srcz=0000000000000000 dstz=0000000000000000 zwdata=0000000000000000 mask=7FFF]
+ [dstart=0 dend=40 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=0 daq=F]
+ Entering A1_ADD state [a1_x=009E, a1_y=0000, addasel=0, addbsel=0, modx=1, addareg=F, adda_xconst=1, adda_yconst=0]...
+ Entering IDLE_INNER state...
+
+ Leaving INNER state... (ocount=0104)
+ [in=T a1f=F a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
+
+ Entering INNER state...
+ Entering DWRITE state... Dest write address/pix address: 000E8508/0 [7730773077307730] (icount=009E, inc=2)
+ srcz=0000000000000000][dcomp=AA zcomp=00 dbinh=00]
+[srcz=0000000000000000 dstz=0000000000000000 zwdata=0000000000000000 mask=7FFF]
+ [dstart=0 dend=40 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=0 daq=F]
+ Entering A1_ADD state [a1_x=00A0, a1_y=0000, addasel=0, addbsel=0, modx=1, addareg=F, adda_xconst=1, adda_yconst=0]...
+ Entering DWRITE state... Dest write address/pix address: 000E8518/0 [7730773077307730] (icount=009C, inc=2)
+ srcz=0000000000000000][dcomp=AA zcomp=00 dbinh=00]
+[srcz=0000000000000000 dstz=0000000000000000 zwdata=0000000000000000 mask=7FFF]
+ [dstart=0 dend=40 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=0 daq=F]
+ Entering A1_ADD state [a1_x=00A2, a1_y=0000, addasel=0, addbsel=0, modx=1, addareg=F, adda_xconst=1, adda_yconst=0]...
+
+*/
+
if (a1_add)
{
#ifdef VERBOSE_BLITTER_LOGGING
fflush(stdout);
}
#endif
-int16 adda_x, adda_y, addb_x, addb_y, data_x, data_y, addq_x, addq_y;
+int16_t adda_x, adda_y, addb_x, addb_y, data_x, data_y, addq_x, addq_y;
ADDAMUX(adda_x, adda_y, addasel, a1_step_x, a1_step_y, a1_stepf_x, a1_stepf_y, a2_step_x, a2_step_y,
a1_inc_x, a1_inc_y, a1_incf_x, a1_incf_y, adda_xconst, adda_yconst, addareg, suba_x, suba_y);
ADDBMUX(addb_x, addb_y, addbsel, a1_x, a1_y, a2_x, a2_y, a1_frac_x, a1_frac_y);
fflush(stdout);
}
#endif
-//void ADDAMUX(int16 &adda_x, int16 &adda_y, uint8 addasel, int16 a1_step_x, int16 a1_step_y,
-// int16 a1_stepf_x, int16 a1_stepf_y, int16 a2_step_x, int16 a2_step_y,
-// int16 a1_inc_x, int16 a1_inc_y, int16 a1_incf_x, int16 a1_incf_y, uint8 adda_xconst,
+//void ADDAMUX(int16_t &adda_x, int16_t &adda_y, uint8_t addasel, int16_t a1_step_x, int16_t a1_step_y,
+// int16_t a1_stepf_x, int16_t a1_stepf_y, int16_t a2_step_x, int16_t a2_step_y,
+// int16_t a1_inc_x, int16_t a1_inc_y, int16_t a1_incf_x, int16_t a1_incf_y, uint8_t adda_xconst,
// bool adda_yconst, bool addareg, bool suba_x, bool suba_y)
-//void ADDBMUX(int16 &addb_x, int16 &addb_y, uint8 addbsel, int16 a1_x, int16 a1_y,
-// int16 a2_x, int16 a2_y, int16 a1_frac_x, int16 a1_frac_y)
-//void ADDRADD(int16 &addq_x, int16 &addq_y, bool a1fracldi,
-// int16 adda_x, int16 adda_y, int16 addb_x, int16 addb_y, uint8 modx, bool suba_x, bool suba_y)
-//void DATAMUX(int16 &data_x, int16 &data_y, uint32 gpu_din, int16 addq_x, int16 addq_y, bool addqsel)
-int16 adda_x, adda_y, addb_x, addb_y, data_x, data_y, addq_x, addq_y;
+//void ADDBMUX(int16_t &addb_x, int16_t &addb_y, uint8_t addbsel, int16_t a1_x, int16_t a1_y,
+// int16_t a2_x, int16_t a2_y, int16_t a1_frac_x, int16_t a1_frac_y)
+//void ADDRADD(int16_t &addq_x, int16_t &addq_y, bool a1fracldi,
+// int16_t adda_x, int16_t adda_y, int16_t addb_x, int16_t addb_y, uint8_t modx, bool suba_x, bool suba_y)
+//void DATAMUX(int16_t &data_x, int16_t &data_y, uint32_t gpu_din, int16_t addq_x, int16_t addq_y, bool addqsel)
+int16_t adda_x, adda_y, addb_x, addb_y, data_x, data_y, addq_x, addq_y;
ADDAMUX(adda_x, adda_y, addasel, a1_step_x, a1_step_y, a1_stepf_x, a1_stepf_y, a2_step_x, a2_step_y,
a1_inc_x, a1_inc_y, a1_incf_x, a1_incf_y, adda_xconst, adda_yconst, addareg, suba_x, suba_y);
ADDBMUX(addb_x, addb_y, addbsel, a1_x, a1_y, a2_x, a2_y, a1_frac_x, a1_frac_y);
//Since we don't get here until the inner loop is finished (indone = true) we can get
//away with doing it here...!
ocount--;
-
+
if (ocount == 0)
outer0 = true;
#ifdef VERBOSE_BLITTER_LOGGING
fflush(stdout);
}
#endif
- uint32 a1_frac_xt = (uint32)a1_frac_x + (uint32)a1_stepf_x;
- uint32 a1_frac_yt = (uint32)a1_frac_y + (uint32)a1_stepf_y;
+ uint32_t a1_frac_xt = (uint32_t)a1_frac_x + (uint32_t)a1_stepf_x;
+ uint32_t a1_frac_yt = (uint32_t)a1_frac_y + (uint32_t)a1_stepf_y;
a1FracCInX = a1_frac_xt >> 16;
a1FracCInY = a1_frac_yt >> 16;
- a1_frac_x = (uint16)(a1_frac_xt & 0xFFFF);
- a1_frac_y = (uint16)(a1_frac_yt & 0xFFFF);
+ a1_frac_x = (uint16_t)(a1_frac_xt & 0xFFFF);
+ a1_frac_y = (uint16_t)(a1_frac_yt & 0xFFFF);
}
if (a1update)
}
#endif
}
-
+
if (a2update)
{
#ifdef VERBOSE_BLITTER_LOGGING
}
}
+// We never get here! !!! FIX !!!
+
+#ifdef VERBOSE_BLITTER_LOGGING
+if (logBlit)
+{
+ printf("Done!\na1_x=%04X a1_y=%04X a1_frac_x=%04X a1_frac_y=%04X a2_x=%04X a2_y%04X\n",
+ GET16(blitter_ram, A1_PIXEL + 2),
+ GET16(blitter_ram, A1_PIXEL + 0),
+ GET16(blitter_ram, A1_FPIXEL + 2),
+ GET16(blitter_ram, A1_FPIXEL + 0),
+ GET16(blitter_ram, A2_PIXEL + 2),
+ GET16(blitter_ram, A2_PIXEL + 0));
+ fflush(stdout);
+}
+#endif
+
// Write values back to registers (in real blitter, these are continuously updated)
- SET16(blitter_ram, A1_PIXEL + 0, a1_y);
SET16(blitter_ram, A1_PIXEL + 2, a1_x);
- SET16(blitter_ram, A1_FPIXEL + 0, a1_frac_y);
+ SET16(blitter_ram, A1_PIXEL + 0, a1_y);
SET16(blitter_ram, A1_FPIXEL + 2, a1_frac_x);
- SET16(blitter_ram, A2_PIXEL + 0, a2_y);
+ SET16(blitter_ram, A1_FPIXEL + 0, a1_frac_y);
SET16(blitter_ram, A2_PIXEL + 2, a2_x);
-}
+ SET16(blitter_ram, A2_PIXEL + 0, a2_y);
+#ifdef VERBOSE_BLITTER_LOGGING
+if (logBlit)
+{
+ printf("Writeback!\na1_x=%04X a1_y=%04X a1_frac_x=%04X a1_frac_y=%04X a2_x=%04X a2_y%04X\n",
+ GET16(blitter_ram, A1_PIXEL + 2),
+ GET16(blitter_ram, A1_PIXEL + 0),
+ GET16(blitter_ram, A1_FPIXEL + 2),
+ GET16(blitter_ram, A1_FPIXEL + 0),
+ GET16(blitter_ram, A2_PIXEL + 2),
+ GET16(blitter_ram, A2_PIXEL + 0));
+ fflush(stdout);
+}
+#endif
+}
/*
-The latest that doesn't work properly:
-
-Blit! (CMD = 09800741)
-Flags: SRCEN CLIP_A1 UPDA1 UPDA1F UPDA2 LFUFUNC=C DCOMPEN
- count = 15 x 18
- a1_base = 00050000, a2_base = 0083F400
- a1_x = 003D, a1_y = 00AD, a1_frac_x = 8000, a1_frac_y = 0000, a2_x = 0027, a2_y = 00A4
- a1_step_x = FFF1, a1_step_y = 0001, a1_stepf_x = 0000, a1_stepf_y = 0000, a2_step_x = FFF1, a2_step_y = 0001
- a1_inc_x = 0001, a1_inc_y = 0000, a1_incf_x = 0000, a1_incf_y = 0000
- a1_win_x = 0180, a1_win_y = 0118, a2_mask_x = 0000, a2_mask_y = 0000
- a2_mask=F a1add=+inc/+0 a2add=+1/+0
- a1_pixsize = 4, a2_pixsize = 4
- srcd=0000000000000000 dstd=0000000000000000 patd=0000000000000000 iinc=00FFF000
- srcz1=0000000000000000 srcz2=0000000000000000 dstz=0000000000000000 zinc=00000000, col=2
- Phrase mode is off
- [in=T a1f=F a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
- Entering INNER state...
- Entering SREAD state... Source read address/pix address: 00858E4E/0 [0000000000000000]
- Entering A2_ADD state [a2_x=0027, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 0007077A/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [0000000000000000] (icount=000E, inc=1)
- Entering A1_ADD state [a1_x=003D, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E50/0 [0000000000000000]
- Entering A2_ADD state [a2_x=0028, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 0007077C/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [0000000000000000] (icount=000D, inc=1)
- Entering A1_ADD state [a1_x=003E, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E52/0 [0000000000000000]
- Entering A2_ADD state [a2_x=0029, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 0007077E/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [0000000000000000] (icount=000C, inc=1)
- Entering A1_ADD state [a1_x=003F, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E54/0 [000000000000014A]
- Entering A2_ADD state [a2_x=002A, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070780/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=000B, inc=1)
- Entering A1_ADD state [a1_x=0040, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E56/0 [000000000000014A]
- Entering A2_ADD state [a2_x=002B, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070782/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=000A, inc=1)
- Entering A1_ADD state [a1_x=0041, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E58/0 [000000000000014A]
- Entering A2_ADD state [a2_x=002C, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070784/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=0009, inc=1)
- Entering A1_ADD state [a1_x=0042, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E5A/0 [000000000000014A]
- Entering A2_ADD state [a2_x=002D, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070786/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=0008, inc=1)
- Entering A1_ADD state [a1_x=0043, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E5C/0 [000000000000014A]
- Entering A2_ADD state [a2_x=002E, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070788/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=0007, inc=1)
- Entering A1_ADD state [a1_x=0044, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E5E/0 [000000000000014A]
- Entering A2_ADD state [a2_x=002F, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 0007078A/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=0006, inc=1)
- Entering A1_ADD state [a1_x=0045, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E60/0 [000000000000014A]
- Entering A2_ADD state [a2_x=0030, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 0007078C/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=0005, inc=1)
- Entering A1_ADD state [a1_x=0046, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E62/0 [000000000000014A]
- Entering A2_ADD state [a2_x=0031, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 0007078E/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=0004, inc=1)
- Entering A1_ADD state [a1_x=0047, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E64/0 [000000000000014A]
- Entering A2_ADD state [a2_x=0032, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070790/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [000000000000014A] (icount=0003, inc=1)
- Entering A1_ADD state [a1_x=0048, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E66/0 [0000000000000000]
- Entering A2_ADD state [a2_x=0033, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070792/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [0000000000000000] (icount=0002, inc=1)
- Entering A1_ADD state [a1_x=0049, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E68/0 [0000000000000000]
- Entering A2_ADD state [a2_x=0034, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070794/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [0000000000000000] (icount=0001, inc=1)
- Entering A1_ADD state [a1_x=004A, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering SREAD state... Source read address/pix address: 00858E6A/0 [0000000000000000]
- Entering A2_ADD state [a2_x=0035, a2_y=00A4, addasel=0, addbsel=1, modx=0, addareg=F, adda_xconst=0, adda_yconst=0]...
- Entering DWRITE state...
- Dest write address/pix address: 00070796/0 [dstart=0 dend=10 pwidth=8 srcshift=0][daas=0 dabs=0 dam=7 ds=1 daq=F] [0000000000000000] (icount=0000, inc=1)
- Entering A1_ADD state [a1_x=004B, a1_y=00AD, addasel=3, addbsel=2, modx=0, addareg=T, adda_xconst=7, adda_yconst=0]...
- Entering IDLE_INNER state...
- Leaving INNER state... (ocount=0011)
- [in=F a1f=T a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
- Entering A1FUPDATE state...
- [in=F a1f=F a1=T zf=F z=F a2=F iif=F iii=F izf=F izi=F]
- Entering A1UPDATE state... (76/173 -> 61/174)
- [in=F a1f=F a1=F zf=F z=F a2=T iif=F iii=F izf=F izi=F]
- Entering A2UPDATE state... (54/164 -> 39/165)
- [in=T a1f=F a1=F zf=F z=F a2=F iif=F iii=F izf=F izi=F]
- Entering INNER state...
+ int16_t a1_x = (int16_t)GET16(blitter_ram, A1_PIXEL + 2);
+ int16_t a1_y = (int16_t)GET16(blitter_ram, A1_PIXEL + 0);
+ uint16_t a1_frac_x = GET16(blitter_ram, A1_FPIXEL + 2);
+ uint16_t a1_frac_y = GET16(blitter_ram, A1_FPIXEL + 0);
+ int16_t a2_x = (int16_t)GET16(blitter_ram, A2_PIXEL + 2);
+ int16_t a2_y = (int16_t)GET16(blitter_ram, A2_PIXEL + 0);
+
+Seems that the ending a1_x should be written between blits, but it doesn't seem to be...
+
+Blit! (CMD = 01800000)
+Flags: LFUFUNC=C
+ count = 28672 x 1
+ a1_base = 00050000, a2_base = 00070000
+ a1_x = 0000, a1_y = 0000, a1_frac_x = 49CD, a1_frac_y = 0000, a2_x = 0033, a2_y = 0001
+ a1_step_x = 0000, a1_step_y = 0000, a1_stepf_x = 939A, a1_stepf_y = 0000, a2_step_x = 0000, a2_step_y = 0000
+ a1_inc_x = 0000, a1_inc_y = 0000, a1_incf_x = 0000, a1_incf_y = 0000
+ a1_win_x = 0100, a1_win_y = 0020, a2_mask_x = 0000, a2_mask_y = 0000
+ a2_mask=F a1add=+phr/+0 a2add=+phr/+0
+ a1_pixsize = 4, a2_pixsize = 3
+ srcd=DEDEDEDEDEDEDEDE dstd=0000000000000000 patd=0000000000000000 iinc=00000000
+ srcz1=0000000000000000 srcz2=0000000000000000 dstz=0000000000000000 zinc=00000000, coll=0
+ Phrase mode is ON
+
+Blit! (CMD = 01800000)
+Flags: LFUFUNC=C
+ count = 28672 x 1
+ a1_base = 00050000, a2_base = 00070000
+ a1_x = 0000, a1_y = 0000, a1_frac_x = 49CD, a1_frac_y = 0000, a2_x = 0033, a2_y = 0001
+ a1_step_x = 0000, a1_step_y = 0000, a1_stepf_x = 939A, a1_stepf_y = 0000, a2_step_x = 0000, a2_step_y = 0000
+ a1_inc_x = 0000, a1_inc_y = 0000, a1_incf_x = 0000, a1_incf_y = 0000
+ a1_win_x = 0100, a1_win_y = 0020, a2_mask_x = 0000, a2_mask_y = 0000
+ a2_mask=F a1add=+phr/+0 a2add=+phr/+0
+ a1_pixsize = 4, a2_pixsize = 3
+ srcd=D6D6D6D6D6D6D6D6 dstd=0000000000000000 patd=0000000000000000 iinc=00000000
+ srcz1=0000000000000000 srcz2=0000000000000000 dstz=0000000000000000 zinc=00000000, coll=0
+ Phrase mode is ON
*/
+
// Various pieces of the blitter puzzle are teased out here...
:IN);
*/
-void ADDRGEN(uint32 &address, uint32 &pixa, bool gena2, bool zaddr,
- uint16 a1_x, uint16 a1_y, uint32 a1_base, uint8 a1_pitch, uint8 a1_pixsize, uint8 a1_width, uint8 a1_zoffset,
- uint16 a2_x, uint16 a2_y, uint32 a2_base, uint8 a2_pitch, uint8 a2_pixsize, uint8 a2_width, uint8 a2_zoffset)
+void ADDRGEN(uint32_t &address, uint32_t &pixa, bool gena2, bool zaddr,
+ uint16_t a1_x, uint16_t a1_y, uint32_t a1_base, uint8_t a1_pitch, uint8_t a1_pixsize, uint8_t a1_width, uint8_t a1_zoffset,
+ uint16_t a2_x, uint16_t a2_y, uint32_t a2_base, uint8_t a2_pitch, uint8_t a2_pixsize, uint8_t a2_width, uint8_t a2_zoffset)
{
- uint16 x = (gena2 ? a2_x : a1_x) & 0x7FFF;
- uint16 y = (gena2 ? a2_y : a1_y) & 0x0FFF;
- uint8 width = (gena2 ? a2_width : a1_width);
- uint8 pixsize = (gena2 ? a2_pixsize : a1_pixsize);
- uint8 pitch = (gena2 ? a2_pitch : a1_pitch);
- uint32 base = (gena2 ? a2_base : a1_base) >> 3;//Only upper 21 bits are passed around the bus? Seems like it...
- uint8 zoffset = (gena2 ? a2_zoffset : a1_zoffset);
+// uint16_t x = (gena2 ? a2_x : a1_x) & 0x7FFF;
+ uint16_t x = (gena2 ? a2_x : a1_x) & 0xFFFF; // Actually uses all 16 bits to generate address...!
+ uint16_t y = (gena2 ? a2_y : a1_y) & 0x0FFF;
+ uint8_t width = (gena2 ? a2_width : a1_width);
+ uint8_t pixsize = (gena2 ? a2_pixsize : a1_pixsize);
+ uint8_t pitch = (gena2 ? a2_pitch : a1_pitch);
+ uint32_t base = (gena2 ? a2_base : a1_base) >> 3;//Only upper 21 bits are passed around the bus? Seems like it...
+ uint8_t zoffset = (gena2 ? a2_zoffset : a1_zoffset);
- uint32 ytm = ((uint32)y << 2) + (width & 0x02 ? (uint32)y << 1 : 0) + (width & 0x01 ? (uint32)y : 0);
+ uint32_t ytm = ((uint32_t)y << 2) + (width & 0x02 ? (uint32_t)y << 1 : 0) + (width & 0x01 ? (uint32_t)y : 0);
- uint32 ya = (ytm << (width >> 2)) >> 2;
+ uint32_t ya = (ytm << (width >> 2)) >> 2;
- uint32 pa = ya + x;
+ uint32_t pa = ya + x;
/*uint32*/ pixa = pa << pixsize;
- uint8 pt = ((pitch & 0x01) && !(pitch & 0x02) ? 0x01 : 0x00)
+ uint8_t pt = ((pitch & 0x01) && !(pitch & 0x02) ? 0x01 : 0x00)
| (!(pitch & 0x01) && (pitch & 0x02) ? 0x02 : 0x00);
-// uint32 phradr = pixa << pt;
- uint32 phradr = (pixa >> 6) << pt;
- uint32 shup = (pitch == 0x03 ? (pixa >> 6) : 0);
+// uint32_t phradr = pixa << pt;
+ uint32_t phradr = (pixa >> 6) << pt;
+ uint32_t shup = (pitch == 0x03 ? (pixa >> 6) : 0);
- uint8 za = (zaddr ? zoffset : 0) & 0x03;
-// uint32 addr = za + (phradr & 0x07) + (shup << 1) + base;
- uint32 addr = za + phradr + (shup << 1) + base;
+ uint8_t za = (zaddr ? zoffset : 0) & 0x03;
+// uint32_t addr = za + (phradr & 0x07) + (shup << 1) + base;
+ uint32_t addr = za + phradr + (shup << 1) + base;
/*uint32*/ address = ((pixa & 0x38) >> 3) | ((addr & 0x1FFFFF) << 3);
#if 0//def VERBOSE_BLITTER_LOGGING
if (logBlit)
INT32/ zstep
:IN);
*/
-void ADDARRAY(uint16 * addq, uint8 daddasel, uint8 daddbsel, uint8 daddmode,
- uint64 dstd, uint32 iinc, uint8 initcin[], uint64 initinc, uint16 initpix,
- uint32 istep, uint64 patd, uint64 srcd, uint64 srcz1, uint64 srcz2,
- uint32 zinc, uint32 zstep)
+void ADDARRAY(uint16_t * addq, uint8_t daddasel, uint8_t daddbsel, uint8_t daddmode,
+ uint64_t dstd, uint32_t iinc, uint8_t initcin[], uint64_t initinc, uint16_t initpix,
+ uint32_t istep, uint64_t patd, uint64_t srcd, uint64_t srcz1, uint64_t srcz2,
+ uint32_t zinc, uint32_t zstep)
{
- uint32 initpix2 = ((uint32)initpix << 16) | initpix;
- uint32 addalo[8], addahi[8];
+ uint32_t initpix2 = ((uint32_t)initpix << 16) | initpix;
+ uint32_t addalo[8], addahi[8];
addalo[0] = dstd & 0xFFFFFFFF;
addalo[1] = initpix2;
addalo[2] = 0;
addahi[5] = patd >> 32;
addahi[6] = srcz1 >> 32;
addahi[7] = srcz2 >> 32;
- uint16 adda[4];
+ uint16_t adda[4];
adda[0] = addalo[daddasel] & 0xFFFF;
adda[1] = addalo[daddasel] >> 16;
adda[2] = addahi[daddasel] & 0xFFFF;
adda[3] = addahi[daddasel] >> 16;
- uint16 wordmux[8];
+ uint16_t wordmux[8];
wordmux[0] = iinc & 0xFFFF;
wordmux[1] = iinc >> 16;
wordmux[2] = zinc & 0xFFFF;
wordmux[5] = istep >> 16;;
wordmux[6] = zstep & 0xFFFF;
wordmux[7] = zstep >> 16;;
- uint16 word = wordmux[((daddbsel & 0x08) >> 1) | (daddbsel & 0x03)];
- uint16 addb[4];
+ uint16_t word = wordmux[((daddbsel & 0x08) >> 1) | (daddbsel & 0x03)];
+ uint16_t addb[4];
bool dbsel2 = daddbsel & 0x04;
bool iincsel = (daddbsel & 0x01) && !(daddbsel & 0x04);
else
addb[0] = addb[1] = addb[2] = addb[3] = 0;
- uint8 cinsel = (daddmode >= 1 && daddmode <= 4 ? 1 : 0);
+ uint8_t cinsel = (daddmode >= 1 && daddmode <= 4 ? 1 : 0);
-static uint8 co[4];//These are preserved between calls...
- uint8 cin[4];
+static uint8_t co[4];//These are preserved between calls...
+ uint8_t cin[4];
for(int i=0; i<4; i++)
cin[i] = initcin[i] | (co[i] & cinsel);
hicinh
:IN);
*/
-void ADD16SAT(uint16 &r, uint8 &co, uint16 a, uint16 b, uint8 cin, bool sat, bool eightbit, bool hicinh)
+void ADD16SAT(uint16_t &r, uint8_t &co, uint16_t a, uint16_t b, uint8_t cin, bool sat, bool eightbit, bool hicinh)
+{
+/*if (logBlit)
{
- uint8 carry[4];
- uint32 qt = (a & 0xFF) + (b & 0xFF) + cin;
+ printf("--> [sat=%s 8b=%s hicinh=%s] %04X + %04X (+ %u) = ", (sat ? "T" : "F"), (eightbit ? "T" : "F"), (hicinh ? "T" : "F"), a, b, cin);
+ fflush(stdout);
+}*/
+ uint8_t carry[4];
+ uint32_t qt = (a & 0xFF) + (b & 0xFF) + cin;
carry[0] = (qt & 0x0100 ? 1 : 0);
- uint16 q = qt & 0x00FF;
+ uint16_t q = qt & 0x00FF;
carry[1] = (carry[0] && !eightbit ? carry[0] : 0);
qt = (a & 0x0F00) + (b & 0x0F00) + (carry[1] << 8);
carry[2] = (qt & 0x1000 ? 1 : 0);
co = (qt & 0x10000 ? 1 : 0);
q |= qt & 0xF000;
- uint8 btop = (eightbit ? (b & 0x0080) >> 7 : (b & 0x8000) >> 15);
- uint8 ctop = (eightbit ? carry[0] : co);
+ uint8_t btop = (eightbit ? (b & 0x0080) >> 7 : (b & 0x8000) >> 15);
+ uint8_t ctop = (eightbit ? carry[0] : co);
bool saturate = sat && (btop ^ ctop);
bool hisaturate = saturate && !eightbit;
+/*if (logBlit)
+{
+ printf("bt=%u ct=%u s=%u hs=%u] ", btop, ctop, saturate, hisaturate);
+ fflush(stdout);
+}*/
r = (saturate ? (ctop ? 0x00FF : 0x0000) : q & 0x00FF);
r |= (hisaturate ? (ctop ? 0xFF00 : 0x0000) : q & 0xFF00);
+/*if (logBlit)
+{
+ printf("%04X (co=%u)\n", r, co);
+ fflush(stdout);
+}*/
}
/** ADDAMUX - Address adder input A selection *******************
adda_xconst[0..2] generate a power of 2 in the range 1-64 or all zeroes when
they are all 1.
-addareg selects register value to be added as opposed to constant
+addareg selects register value to be added as opposed to constant
value.
suba_x, suba_y complement the X and Y values
suba_x
suba_y :IN);
*/
-void ADDAMUX(int16 &adda_x, int16 &adda_y, uint8 addasel, int16 a1_step_x, int16 a1_step_y,
- int16 a1_stepf_x, int16 a1_stepf_y, int16 a2_step_x, int16 a2_step_y,
- int16 a1_inc_x, int16 a1_inc_y, int16 a1_incf_x, int16 a1_incf_y, uint8 adda_xconst,
+void ADDAMUX(int16_t &adda_x, int16_t &adda_y, uint8_t addasel, int16_t a1_step_x, int16_t a1_step_y,
+ int16_t a1_stepf_x, int16_t a1_stepf_y, int16_t a2_step_x, int16_t a2_step_y,
+ int16_t a1_inc_x, int16_t a1_inc_y, int16_t a1_incf_x, int16_t a1_incf_y, uint8_t adda_xconst,
bool adda_yconst, bool addareg, bool suba_x, bool suba_y)
{
-/*INT16/ addac_x, addac_y, addar_x, addar_y, addart_x, addart_y,
+/*INT16/ addac_x, addac_y, addar_x, addar_y, addart_x, addart_y,
INT16/ addas_x, addas_y, suba_x16, suba_y16
:LOCAL;
BEGIN
Addar_y := MX2 (addar_y, addart_y, a2_step_y, addaselb[2]);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- int16 xterm[4], yterm[4];
+ int16_t xterm[4], yterm[4];
xterm[0] = a1_step_x, xterm[1] = a1_stepf_x, xterm[2] = a1_inc_x, xterm[3] = a1_incf_x;
yterm[0] = a1_step_y, yterm[1] = a1_stepf_y, yterm[2] = a1_inc_y, yterm[3] = a1_incf_y;
- int16 addar_x = (addasel & 0x04 ? a2_step_x : xterm[addasel & 0x03]);
- int16 addar_y = (addasel & 0x04 ? a2_step_y : yterm[addasel & 0x03]);
+ int16_t addar_x = (addasel & 0x04 ? a2_step_x : xterm[addasel & 0x03]);
+ int16_t addar_y = (addasel & 0x04 ? a2_step_y : yterm[addasel & 0x03]);
//////////////////////////////////////////////////////////////////////////////////////
-/* Generate a constant value - this is a power of 2 in the range
+/* Generate a constant value - this is a power of 2 in the range
0-64, or zero. The control bits are adda_xconst[0..2], when they
are all 1 the result is 0.
Constants for Y can only be 0 or 1 */
Unused[0] := DUMMY (unused[0]);
Addac_x := JOIN (addac_x, addac_x[0..6], zero, zero, zero, zero, zero, zero, zero, zero, zero);
-Addac_y := JOIN (addac_y, adda_yconst, zero, zero, zero, zero, zero, zero, zero, zero, zero, zero,
+Addac_y := JOIN (addac_y, adda_yconst, zero, zero, zero, zero, zero, zero, zero, zero, zero, zero,
zero, zero, zero, zero, zero);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- int16 addac_x = (adda_xconst == 0x07 ? 0 : 1 << adda_xconst);
- int16 addac_y = (adda_yconst ? 0x01 : 0);
+ int16_t addac_x = (adda_xconst == 0x07 ? 0 : 1 << adda_xconst);
+ int16_t addac_y = (adda_yconst ? 0x01 : 0);
//////////////////////////////////////////////////////////////////////////////////////
/* Select between constant value and register value */
/*Addas_x := MX2 (addas_x, addac_x, addar_x, addareg);
Addas_y := MX2 (addas_y, addac_y, addar_y, addareg);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- int16 addas_x = (addareg ? addar_x : addac_x);
- int16 addas_y = (addareg ? addar_y : addac_y);
+ int16_t addas_x = (addareg ? addar_x : addac_x);
+ int16_t addas_y = (addareg ? addar_y : addac_y);
//////////////////////////////////////////////////////////////////////////////////////
/* Complement these values (complement flag gives adder carry in)*/
-/*Suba_x16 := JOIN (suba_x16, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x,
+/*Suba_x16 := JOIN (suba_x16, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x,
suba_x, suba_x, suba_x, suba_x, suba_x, suba_x, suba_x);
-Suba_y16 := JOIN (suba_y16, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y,
+Suba_y16 := JOIN (suba_y16, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y,
suba_y, suba_y, suba_y, suba_y, suba_y, suba_y, suba_y);
Adda_x := EO (adda_x, suba_x16, addas_x);
Adda_y := EO (adda_y, suba_y16, addas_y);*/
/** ADDBMUX - Address adder input B selection *******************
-This module selects the register to be updated by the address
-adder. This can be one of three registers, the A1 and A2
-pointers, or the A1 fractional part. It can also be zero, so that the step
+This module selects the register to be updated by the address
+adder. This can be one of three registers, the A1 and A2
+pointers, or the A1 fractional part. It can also be zero, so that the step
registers load directly into the pointers.
*/
:IN);
INT16/ zero16 :LOCAL;
BEGIN*/
-void ADDBMUX(int16 &addb_x, int16 &addb_y, uint8 addbsel, int16 a1_x, int16 a1_y,
- int16 a2_x, int16 a2_y, int16 a1_frac_x, int16 a1_frac_y)
+void ADDBMUX(int16_t &addb_x, int16_t &addb_y, uint8_t addbsel, int16_t a1_x, int16_t a1_y,
+ int16_t a2_x, int16_t a2_y, int16_t a1_frac_x, int16_t a1_frac_y)
{
/*Zero := TIE0 (zero);
-Zero16 := JOIN (zero16, zero, zero, zero, zero, zero, zero, zero,
+Zero16 := JOIN (zero16, zero, zero, zero, zero, zero, zero, zero,
zero, zero, zero, zero, zero, zero, zero, zero, zero);
Addbselb[0-1] := BUF8 (addbselb[0-1], addbsel[0-1]);
Addb_x := MX4 (addb_x, a1_x, a2_x, a1_frac_x, zero16, addbselb[0..1]);
Addb_y := MX4 (addb_y, a1_y, a2_y, a1_frac_y, zero16, addbselb[0..1]);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- int16 xterm[4], yterm[4];
+ int16_t xterm[4], yterm[4];
xterm[0] = a1_x, xterm[1] = a2_x, xterm[2] = a1_frac_x, xterm[3] = 0;
yterm[0] = a1_y, yterm[1] = a2_y, yterm[2] = a1_frac_y, yterm[3] = 0;
addb_x = xterm[addbsel & 0x03];
INT16/ gpu_lo, gpu_hi
:LOCAL;
BEGIN*/
-void DATAMUX(int16 &data_x, int16 &data_y, uint32 gpu_din, int16 addq_x, int16 addq_y, bool addqsel)
+void DATAMUX(int16_t &data_x, int16_t &data_y, uint32_t gpu_din, int16_t addq_x, int16_t addq_y, bool addqsel)
{
/*Gpu_lo := JOIN (gpu_lo, gpu_din{0..15});
Gpu_hi := JOIN (gpu_hi, gpu_din{16..31});
Data_x := MX2 (data_x, gpu_lo, addq_x, addqselb);
Data_y := MX2 (data_y, gpu_hi, addq_y, addqselb);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- data_x = (addqsel ? addq_x : (int16)(gpu_din & 0xFFFF));
- data_y = (addqsel ? addq_y : (int16)(gpu_din >> 16));
+ data_x = (addqsel ? addq_x : (int16_t)(gpu_din & 0xFFFF));
+ data_y = (addqsel ? addq_y : (int16_t)(gpu_din >> 16));
//////////////////////////////////////////////////////////////////////////////////////
//END;
Blitter Address Adder
---------------------
-The blitter address adder is a pair of sixteen bit adders, one
-each for X and Y. The multiplexing of the input terms is
-performed elsewhere, but this adder can also perform modulo
-arithmetic to align X-addresses onto phrase boundaries.
+The blitter address adder is a pair of sixteen bit adders, one
+each for X and Y. The multiplexing of the input terms is
+performed elsewhere, but this adder can also perform modulo
+arithmetic to align X-addresses onto phrase boundaries.
modx[0..2] take values
000 no mask
modx[0..2]
suba_x
suba_y
- :IN);
+ :IN);
BEGIN
Zero := TIE0 (zero);*/
-void ADDRADD(int16 &addq_x, int16 &addq_y, bool a1fracldi,
- uint16 adda_x, uint16 adda_y, uint16 addb_x, uint16 addb_y, uint8 modx, bool suba_x, bool suba_y)
+void ADDRADD(int16_t &addq_x, int16_t &addq_y, bool a1fracldi,
+ uint16_t adda_x, uint16_t adda_y, uint16_t addb_x, uint16_t addb_y, uint8_t modx, bool suba_x, bool suba_y)
{
/* Perform the addition */
////////////////////////////////////// C++ CODE //////////////////////////////////////
//I'm sure the following will generate a bunch of warnings, but will have to do for now.
- static uint16 co_x = 0, co_y = 0; // Carry out has to propogate between function calls...
- uint16 ci_x = co_x ^ (suba_x ? 1 : 0);
- uint16 ci_y = co_y ^ (suba_y ? 1 : 0);
- uint32 addqt_x = adda_x + addb_x + ci_x;
- uint32 addqt_y = adda_y + addb_y + ci_y;
+ static uint16_t co_x = 0, co_y = 0; // Carry out has to propogate between function calls...
+ uint16_t ci_x = co_x ^ (suba_x ? 1 : 0);
+ uint16_t ci_y = co_y ^ (suba_y ? 1 : 0);
+ uint32_t addqt_x = adda_x + addb_x + ci_x;
+ uint32_t addqt_y = adda_y + addb_y + ci_y;
co_x = ((addqt_x & 0x10000) && a1fracldi ? 1 : 0);
co_y = ((addqt_y & 0x10000) && a1fracldi ? 1 : 0);
//////////////////////////////////////////////////////////////////////////////////////
Addq_y := JOIN (addq_y, addq_y[0..15]);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- int16 mask[8] = { 0xFFFF, 0xFFFE, 0xFFFC, 0xFFF8, 0xFFF0, 0xFFE0, 0xFFC0, 0x0000 };
+ int16_t mask[8] = { 0xFFFF, 0xFFFE, 0xFFFC, 0xFFF8, 0xFFF0, 0xFFE0, 0xFFC0, 0x0000 };
addq_x = addqt_x & mask[modx];
addq_y = addqt_y & 0xFFFF;
//////////////////////////////////////////////////////////////////////////////////////
dcomp[0..7] // data byte equal flags
srcd[0..7] // bits to use for bit to byte expansion
zcomp[0..3] // output from Z comparators
- :OUT;
+ :OUT;
a1_x[0..1] // low two bits of A1 X pointer
big_pix // pixel organisation is big-endian
blitter_active // blitter is active
:IN);
*/
-void DATA (uint64 &wdata, uint8 &dcomp, uint8 &zcomp, bool &nowrite,
- bool big_pix, bool cmpdst, uint8 daddasel, uint8 daddbsel, uint8 daddmode, bool daddq_sel, uint8 data_sel,
- uint8 dbinh, uint8 dend, uint8 dstart, uint64 dstd, uint32 iinc, uint8 lfu_func, uint64 patd, bool patdadd,
- bool phrase_mode, uint64 srcd, bool srcdread, bool srczread, bool srcz2add, uint8 zmode,
- bool bcompen, bool bkgwren, bool dcompen, uint8 icount, uint8 pixsize)
+void DATA(uint64_t &wdata, uint8_t &dcomp, uint8_t &zcomp, bool &nowrite,
+ bool big_pix, bool cmpdst, uint8_t daddasel, uint8_t daddbsel, uint8_t daddmode, bool daddq_sel, uint8_t data_sel,
+ uint8_t dbinh, uint8_t dend, uint8_t dstart, uint64_t dstd, uint32_t iinc, uint8_t lfu_func, uint64_t &patd, bool patdadd,
+ bool phrase_mode, uint64_t srcd, bool srcdread, bool srczread, bool srcz2add, uint8_t zmode,
+ bool bcompen, bool bkgwren, bool dcompen, uint8_t icount, uint8_t pixsize,
+ uint64_t &srcz, uint64_t dstz, uint32_t zinc)
{
/*
Stuff we absolutely *need* to have passed in/out:
IN:
patdadd, dstd, srcd, patd, daddasel, daddbsel, daddmode, iinc, srcz1, srcz2, big_pix, phrase_mode, cmpdst
OUT:
- changed patd (wdata I guess...)
+ changed patd (wdata I guess...) (Nope. We pass it back directly now...)
*/
// Source data registers
Srcd[8-31] := JOIN (srcd[8-31], srcdlo{8-31});
Srcd[32-63] := JOIN (srcd[32-63], srcdhi{0-31});*/
-// Destination data registers
+// Destination data registers
/*Data_dst := DATA_DST (dstd[0..63], dstz[0..1], clk, dstdld[0..1], dstzld[0..1], load_data[0..1]);
Dstdlo := JOIN (dstdlo, dstd[0..31]);
Dstdhi := JOIN (dstdhi, dstd[32..63]);*/
-// Pattern and Color data registers
+// Pattern and Color data registers
// Looks like this is simply another register file for the pattern data registers. No adding or anything funky
// going on. Note that patd & patdv will output the same info.
// Multiplying data Mixer (NOT IN JAGUAR I)
-/*Datamix := DATAMIX (patdo[0..1], clk, colord[0..15], dpipe[1], dstd[0..63], int0dp[8..10], int1dp[8..10],
+/*Datamix := DATAMIX (patdo[0..1], clk, colord[0..15], dpipe[1], dstd[0..63], int0dp[8..10], int1dp[8..10],
int2dp[8..10], int3dp[8..10], mixsel[0..2], patd[0..63], pdsel[0..1], srcd[0..63], textrgb, txtd[0..63]);*/
// Logic function unit
/*Lfu := LFU (lfu[0..1], srcdlo, srcdhi, dstdlo, dstdhi, lfu_func[0..3]);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- uint64 funcmask[2] = { 0, 0xFFFFFFFFFFFFFFFFLL };
- uint64 func0 = funcmask[lfu_func & 0x01];
- uint64 func1 = funcmask[(lfu_func >> 1) & 0x01];
- uint64 func2 = funcmask[(lfu_func >> 2) & 0x01];
- uint64 func3 = funcmask[(lfu_func >> 3) & 0x01];
- uint64 lfu = (~srcd & ~dstd & func0) | (~srcd & dstd & func1) | (srcd & ~dstd & func2) | (srcd & dstd & func3);
+ uint64_t funcmask[2] = { 0, 0xFFFFFFFFFFFFFFFFLL };
+ uint64_t func0 = funcmask[lfu_func & 0x01];
+ uint64_t func1 = funcmask[(lfu_func >> 1) & 0x01];
+ uint64_t func2 = funcmask[(lfu_func >> 2) & 0x01];
+ uint64_t func3 = funcmask[(lfu_func >> 3) & 0x01];
+ uint64_t lfu = (~srcd & ~dstd & func0) | (~srcd & dstd & func1) | (srcd & ~dstd & func2) | (srcd & dstd & func3);
//////////////////////////////////////////////////////////////////////////////////////
// Increment and Step Registers
-
+
// Does it do anything without the step add lines? Check it!
// No. This is pretty much just a register file without the Jaguar II lines...
/*Inc_step := INC_STEP (iinc, istep[0..31], zinc, zstep[0..31], clk, ext_int, gpu_din, iincld, iincldx, istepadd,
/*Datacomp := DATACOMP (dcomp[0..7], cmpdst, dstdlo, dstdhi, patdlo, patdhi, srcdlo, srcdhi);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
dcomp = 0;
- uint64 cmpd = patd ^ (cmpdst ? dstd : srcd);
+ uint64_t cmpd = patd ^ (cmpdst ? dstd : srcd);
if ((cmpd & 0x00000000000000FFLL) == 0)
dcomp |= 0x01;
dcomp |= 0x40;
if ((cmpd & 0xFF00000000000000LL) == 0)
dcomp |= 0x80;
+//////////////////////////////////////////////////////////////////////////////////////
+
+// Zed comparator for Z-buffer operations
+
+/*Zedcomp := ZEDCOMP (zcomp[0..3], srczp[0..1], dstz[0..1], zmode[0..2]);*/
+////////////////////////////////////// C++ CODE //////////////////////////////////////
+//srczp is srcz pipelined, also it goes through a source shift as well...
+/*The shift is basically like so (each piece is 16 bits long):
+
+ 0 1 2 3 4 5 6
+ srcz1lolo srcz1lohi srcz1hilo srcz1hihi srcrz2lolo srcz2lohi srcz2hilo
+
+with srcshift bits 4 & 5 selecting the start position
+*/
+//So... basically what we have here is:
+ zcomp = 0;
+
+ if ((((srcz & 0x000000000000FFFFLL) < (dstz & 0x000000000000FFFFLL)) && (zmode & 0x01))
+ || (((srcz & 0x000000000000FFFFLL) == (dstz & 0x000000000000FFFFLL)) && (zmode & 0x02))
+ || (((srcz & 0x000000000000FFFFLL) > (dstz & 0x000000000000FFFFLL)) && (zmode & 0x04)))
+ zcomp |= 0x01;
+
+ if ((((srcz & 0x00000000FFFF0000LL) < (dstz & 0x00000000FFFF0000LL)) && (zmode & 0x01))
+ || (((srcz & 0x00000000FFFF0000LL) == (dstz & 0x00000000FFFF0000LL)) && (zmode & 0x02))
+ || (((srcz & 0x00000000FFFF0000LL) > (dstz & 0x00000000FFFF0000LL)) && (zmode & 0x04)))
+ zcomp |= 0x02;
+
+ if ((((srcz & 0x0000FFFF00000000LL) < (dstz & 0x0000FFFF00000000LL)) && (zmode & 0x01))
+ || (((srcz & 0x0000FFFF00000000LL) == (dstz & 0x0000FFFF00000000LL)) && (zmode & 0x02))
+ || (((srcz & 0x0000FFFF00000000LL) > (dstz & 0x0000FFFF00000000LL)) && (zmode & 0x04)))
+ zcomp |= 0x04;
+
+ if ((((srcz & 0xFFFF000000000000LL) < (dstz & 0xFFFF000000000000LL)) && (zmode & 0x01))
+ || (((srcz & 0xFFFF000000000000LL) == (dstz & 0xFFFF000000000000LL)) && (zmode & 0x02))
+ || (((srcz & 0xFFFF000000000000LL) > (dstz & 0xFFFF000000000000LL)) && (zmode & 0x04)))
+ zcomp |= 0x08;
+
+//TEMP, TO TEST IF ZCOMP IS THE CULPRIT...
+//Nope, this is NOT the problem...
+//zcomp=0;
// We'll do the comparison/bit/byte inhibits here, since that's they way it happens
// in the real thing (dcomp goes out to COMP_CTRL and back into DATA through dbinh)...
#if 1
- uint8 dbinht;
+ uint8_t dbinht;
// bool nowrite;
COMP_CTRL(dbinht, nowrite,
- bcompen, true/*big_pix*/, bkgwren, dcomp, dcompen, icount, pixsize, phrase_mode, srcd & 0xFF, 0);//zcomp);
+ bcompen, true/*big_pix*/, bkgwren, dcomp, dcompen, icount, pixsize, phrase_mode, srcd & 0xFF, zcomp);
dbinh = dbinht;
// dbinh = 0x00;
#endif
-//////////////////////////////////////////////////////////////////////////////////////
-
-// Zed comparator for Z-buffer operations
-
-/*Zedcomp := ZEDCOMP (zcomp[0..3], srczp[0..1], dstz[0..1], zmode[0..2]);*/
-////////////////////////////////////// C++ CODE //////////////////////////////////////
+#if 1
+#ifdef VERBOSE_BLITTER_LOGGING
+if (logBlit)
+{
+ printf("\n[dcomp=%02X zcomp=%02X dbinh=%02X]\n", dcomp, zcomp, dbinh);
+ fflush(stdout);
+}//*/
+#endif
+#endif
//////////////////////////////////////////////////////////////////////////////////////
// 22 Mar 94
// The data initializer - allows all four initial values to be computed from one (NOT IN JAGUAR I)
-/*Datinit := DATINIT (initcin[0..3], initinc[0..63], initpix[0..15], a1_x[0..1], big_pix, clk, iinc, init_if, init_ii,
+/*Datinit := DATINIT (initcin[0..3], initinc[0..63], initpix[0..15], a1_x[0..1], big_pix, clk, iinc, init_if, init_ii,
init_zf, istep[0..31], zinc, zstep[0..31]);*/
// Adder array for Z and intensity increments
/*Addarray := ADDARRAY (addq[0..3], clk, daddasel[0..2], daddbsel[0..3], daddmode[0..2], dstdlo, dstdhi, iinc,
initcin[0..3], initinc[0..63], initpix[0..15], istep, patdv[0..1], srcdlo, srcdhi, srcz1[0..1],
srcz2[0..1], reset\, zinc, zstep);*/
-/*void ADDARRAY(uint16 * addq, uint8 daddasel, uint8 daddbsel, uint8 daddmode,
- uint64 dstd, uint32 iinc, uint8 initcin[], uint64 initinc, uint16 initpix,
- uint32 istep, uint64 patd, uint64 srcd, uint64 srcz1, uint64 srcz2,
- uint32 zinc, uint32 zstep)*/
+/*void ADDARRAY(uint16_t * addq, uint8_t daddasel, uint8_t daddbsel, uint8_t daddmode,
+ uint64_t dstd, uint32_t iinc, uint8_t initcin[], uint64_t initinc, uint16_t initpix,
+ uint32_t istep, uint64_t patd, uint64_t srcd, uint64_t srcz1, uint64_t srcz2,
+ uint32_t zinc, uint32_t zstep)*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- uint16 addq[4];
- uint8 initcin[4] = { 0, 0, 0, 0 };
+ uint16_t addq[4];
+ uint8_t initcin[4] = { 0, 0, 0, 0 };
ADDARRAY(addq, daddasel, daddbsel, daddmode, dstd, iinc, initcin, 0, 0, 0, patd, srcd, 0, 0, 0, 0);
//This is normally done asynchronously above (thru local_data) when in patdadd mode...
+//And now it's passed back to the caller to be persistent between calls...!
+//But it's causing some serious fuck-ups in T2K now... !!! FIX !!! [DONE--???]
+//Weird! It doesn't anymore...!
if (patdadd)
- patd = ((uint64)addq[3] << 48) | ((uint64)addq[2] << 32) | ((uint64)addq[1] << 16) | (uint64)addq[0];
+ patd = ((uint64_t)addq[3] << 48) | ((uint64_t)addq[2] << 32) | ((uint64_t)addq[1] << 16) | (uint64_t)addq[0];
//////////////////////////////////////////////////////////////////////////////////////
// Local data bus multiplexer
E_coarse[0] := INV1 (e_coarse[0], e_coarse\[0]);
Efine := DECL38E (unused[0], e_fine\[1..7], dend[0..2], e_coarse[0]);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- uint8 decl38e[2][8] = { { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+ uint8_t decl38e[2][8] = { { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
{ 0xFE, 0xFD, 0xFB, 0xF7, 0xEF, 0xDF, 0xBF, 0x7F } };
- uint8 dech38[8] = { 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80 };
- uint8 dech38el[2][8] = { { 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80 },
- { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF } };
+ uint8_t dech38[8] = { 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80 };
+ uint8_t dech38el[2][8] = { { 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80 },
+ { 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 } };
int en = (dend & 0x3F ? 1 : 0);
- uint8 e_coarse = decl38e[en][(dend & 0x38) >> 3]; // Actually, this is e_coarse inverted...
- uint8 e_fine = decl38e[(e_coarse & 0x01) ^ 0x01][dend & 0x07];
+ uint8_t e_coarse = decl38e[en][(dend & 0x38) >> 3]; // Actually, this is e_coarse inverted...
+ uint8_t e_fine = decl38e[(e_coarse & 0x01) ^ 0x01][dend & 0x07];
e_fine &= 0xFE;
//////////////////////////////////////////////////////////////////////////////////////
Sfen\ := INV1 (sfen\, s_coarse[0]);
Sfine := DECH38EL (s_fine[0..7], dstart[0..2], sfen\);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- uint8 s_coarse = dech38[(dstart & 0x38) >> 3];
- uint8 s_fine = dech38el[(s_coarse & 0x01) ^ 0x01][dstart & 0x07];
+ uint8_t s_coarse = dech38[(dstart & 0x38) >> 3];
+ uint8_t s_fine = dech38el[(s_coarse & 0x01) ^ 0x01][dstart & 0x07];
//////////////////////////////////////////////////////////////////////////////////////
/*Maskt[0] := BUF1 (maskt[0], s_fine[0]);
Maskt[1-7] := OAN1P (maskt[1-7], maskt[0-6], s_fine[1-7], e_fine\[1-7]);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- uint16 maskt = s_fine & 0x0001;
+ uint16_t maskt = s_fine & 0x0001;
maskt |= (((maskt & 0x0001) || (s_fine & 0x02)) && (e_fine & 0x02) ? 0x0002 : 0x0000);
maskt |= (((maskt & 0x0002) || (s_fine & 0x04)) && (e_fine & 0x04) ? 0x0004 : 0x0000);
maskt |= (((maskt & 0x0004) || (s_fine & 0x08)) && (e_fine & 0x08) ? 0x0008 : 0x0000);
//////////////////////////////////////////////////////////////////////////////////////
/* The bit terms are mirrored for big-endian pixels outside phrase
-mode. The byte terms are mirrored for big-endian pixels in phrase
+mode. The byte terms are mirrored for big-endian pixels in phrase
mode. */
/*Mirror_bit := AN2M (mir_bit, phrase_mode\, big_pix);
////////////////////////////////////// C++ CODE //////////////////////////////////////
bool mir_bit = true/*big_pix*/ && !phrase_mode;
bool mir_byte = true/*big_pix*/ && phrase_mode;
- uint16 masku = maskt;
+ uint16_t masku = maskt;
if (mir_bit)
{
/*Mask[0-7] := AN2 (mask[0-7], masku[0-7], dbinh\[0]);
Mask[8-14] := AN2H (mask[8-14], masku[8-14], dbinh\[1-7]);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- uint16 mask = masku & (!(dbinh & 0x01) ? 0xFFFF : 0xFF00);
- mask &= ~(((uint16)dbinh & 0x00FE) << 7);
+ uint16_t mask = masku & (!(dbinh & 0x01) ? 0xFFFF : 0xFF00);
+ mask &= ~(((uint16_t)dbinh & 0x00FE) << 7);
//////////////////////////////////////////////////////////////////////////////////////
/*Addql[0] := JOIN (addql[0], addq[0..1]);
Ddatlo := MX4 (ddatlo, patd[0], lfu[0], addql[0], zero32, dsel0b[0], dsel1b[0]);
Ddathi := MX4 (ddathi, patd[1], lfu[1], addql[1], zero32, dsel0b[1], dsel1b[1]);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
- uint64 dmux[4];
+ uint64_t dmux[4];
dmux[0] = patd;
dmux[1] = lfu;
- dmux[2] = ((uint64)addq[3] << 48) | ((uint64)addq[2] << 32) | ((uint64)addq[1] << 16) | (uint64)addq[0];
+ dmux[2] = ((uint64_t)addq[3] << 48) | ((uint64_t)addq[2] << 32) | ((uint64_t)addq[1] << 16) | (uint64_t)addq[0];
dmux[3] = 0;
- uint64 ddat = dmux[data_sel];
+ uint64_t ddat = dmux[data_sel];
//////////////////////////////////////////////////////////////////////////////////////
/*Zed_sel := AN2 (zed_sel, data_sel[0..1]);
wdata |= (mask & 0x1000 ? ddat : dstd) & 0x0000FF0000000000LL;
wdata |= (mask & 0x2000 ? ddat : dstd) & 0x00FF000000000000LL;
wdata |= (mask & 0x4000 ? ddat : dstd) & 0xFF00000000000000LL;
+/*if (logBlit)
+{
+ printf("\n[ddat=%08X%08X dstd=%08X%08X wdata=%08X%08X mask=%04X]\n",
+ (uint32_t)(ddat >> 32), (uint32_t)(ddat & 0xFFFFFFFF),
+ (uint32_t)(dstd >> 32), (uint32_t)(dstd & 0xFFFFFFFF),
+ (uint32_t)(wdata >> 32), (uint32_t)(wdata & 0xFFFFFFFF), mask);
+ fflush(stdout);
+}//*/
+//This is a crappy way of handling this, but it should work for now...
+ uint64_t zwdata;
+ zwdata = ((srcz & mask) | (dstz & ~mask)) & 0x00000000000000FFLL;
+ zwdata |= (mask & 0x0100 ? srcz : dstz) & 0x000000000000FF00LL;
+ zwdata |= (mask & 0x0200 ? srcz : dstz) & 0x0000000000FF0000LL;
+ zwdata |= (mask & 0x0400 ? srcz : dstz) & 0x00000000FF000000LL;
+ zwdata |= (mask & 0x0800 ? srcz : dstz) & 0x000000FF00000000LL;
+ zwdata |= (mask & 0x1000 ? srcz : dstz) & 0x0000FF0000000000LL;
+ zwdata |= (mask & 0x2000 ? srcz : dstz) & 0x00FF000000000000LL;
+ zwdata |= (mask & 0x4000 ? srcz : dstz) & 0xFF00000000000000LL;
+if (logBlit)
+{
+ printf("\n[srcz=%08X%08X dstz=%08X%08X zwdata=%08X%08X mask=%04X]\n",
+ (uint32_t)(srcz >> 32), (uint32_t)(srcz & 0xFFFFFFFF),
+ (uint32_t)(dstz >> 32), (uint32_t)(dstz & 0xFFFFFFFF),
+ (uint32_t)(zwdata >> 32), (uint32_t)(zwdata & 0xFFFFFFFF), mask);
+ fflush(stdout);
+}//*/
+ srcz = zwdata;
//////////////////////////////////////////////////////////////////////////////////////
/*Data_enab[0-1] := BUF8 (data_enab[0-1], data_ena);
/** COMP_CTRL - Comparator output control logic *****************
This block is responsible for taking the comparator outputs and
-using them as appropriate to inhibit writes. Two methods are
+using them as appropriate to inhibit writes. Two methods are
supported for inhibiting write data:
- suppression of the inner loop controlled write operation
Writes can be suppressed by data being equal, by the Z comparator
conditions being met, or by the bit to pixel expansion scheme.
-Pipe-lining issues: the data derived comparator outputs are stable
+Pipe-lining issues: the data derived comparator outputs are stable
until the next data read, well after the affected write from this
operation. However, the inner counter bits can count immediately
-before the ack for the last write. Therefore, it is necessary to
+before the ack for the last write. Therefore, it is necessary to
delay bcompbit select terms by one inner loop pipe-line stage,
when generating the select for the data control - the output is
-delayed one further tick to give it write data timing (2/34).
+delayed one further tick to give it write data timing (2/34).
There is also a problem with computed data - the new values are
calculated before the write associated with the old value has been
step_inner // inner loop advance
zcomp[0..3] // output of word zed comparators
:IN);*/
-void COMP_CTRL(uint8 &dbinh, bool &nowrite,
- bool bcompen, bool big_pix, bool bkgwren, uint8 dcomp, bool dcompen, uint8 icount,
- uint8 pixsize, bool phrase_mode, uint8 srcd, uint8 zcomp)
+void COMP_CTRL(uint8_t &dbinh, bool &nowrite,
+ bool bcompen, bool big_pix, bool bkgwren, uint8_t dcomp, bool dcompen, uint8_t icount,
+ uint8_t pixsize, bool phrase_mode, uint8_t srcd, uint8_t zcomp)
{
//BEGIN
only applicable to 8-bit pixel mode (2/34) */
/*Bcompselt[0-2] := EO (bcompselt[0-2], icount[0-2], big_pix);
-Bcompbit := MX8 (bcompbit, srcd[7], srcd[6], srcd[5],
+Bcompbit := MX8 (bcompbit, srcd[7], srcd[6], srcd[5],
srcd[4], srcd[3], srcd[2], srcd[1], srcd[0], bcompselt[0..2]);
Bcompbit\ := INV1 (bcompbit\, bcompbit);*/
////////////////////////////////////// C++ CODE //////////////////////////////////////
fflush(stdout);
}
#endif
- uint8 bcompselt = (big_pix ? ~icount : icount) & 0x07;
- uint8 bitmask[8] = { 0x80, 0x40, 0x20, 0x10, 0x08, 0x04, 0x02, 0x01 };
+ uint8_t bcompselt = (big_pix ? ~icount : icount) & 0x07;
+ uint8_t bitmask[8] = { 0x80, 0x40, 0x20, 0x10, 0x08, 0x04, 0x02, 0x01 };
bool bcompbit = srcd & bitmask[bcompselt];
//////////////////////////////////////////////////////////////////////////////////////
/* pipe-line the count */
/*Bcompsel[0-2] := FDSYNC (bcompsel[0-2], bcompselt[0-2], step_inner, clk);
-Bcompbt := MX8 (bcompbitpt, srcd[7], srcd[6], srcd[5],
+Bcompbt := MX8 (bcompbitpt, srcd[7], srcd[6], srcd[5],
srcd[4], srcd[3], srcd[2], srcd[1], srcd[0], bcompsel[0..2]);
Bcompbitp := FD1Q (bcompbitp, bcompbitpt, clk);
Bcompbitp\ := INV1 (bcompbitp\, bcompbitp);*/
/* For pixel mode, generate the write inhibit signal for all modes
on bit inhibit, for 8 and 16 bit modes on comparator inhibit, and
-for 16 bit mode on Z inhibit
+for 16 bit mode on Z inhibit
Nowrite = bcompen . /bcompbit . /phrase_mode
+ dcompen . dcomp[0] . /phrase_mode . pixsize = 011
//////////////////////////////////////////////////////////////////////////////////////
/* For phrase mode, generate the byte inhibit signals for eight bit
-mode 011, or sixteen bit mode 100
+mode 011, or sixteen bit mode 100
dbinh\[0] = pixsize[2] . zcomp[0]
+ pixsize[2] . dcomp[0] . dcomp[1] . dcompen
+ /pixsize[2] . dcomp[0] . dcompen
////////////////////////////////////// C++ CODE //////////////////////////////////////
//////////////////////////////////////////////////////////////////////////////////////
+// !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!!
+// !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!!
+// !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!! TESTING !!!
+
#endif
+
projects / virtualjaguar / blobdiff