//
// Blitter core
//
-// by cal2
+// by Cal2
// GCC/SDL port by Niels Wagenaar (Linux/WIN32) and Caz (BeOS)
// Cleanups/fixes by James L. Hammons
//
+bool specialLog = false;
+extern int effect_start;
+extern int blit_start_log;
-//#define GEN_CODE
-//#define LOG_BLITS
-// Generic blitter it is, until all blitter bugs are squashed!
-#define USE_GENERIC_BLITTER
-
#include "jaguar.h"
-#include "blitter2.h" // Testing purposes only!
-
-#define null 0
-extern int jaguar_active_memory_dumps;
#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 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)
-int start_logging = 0;
-
-// Blitter register RAM (most of it is hidden from the user)
-
-static uint8 blitter_ram[0x100];
-
// Blitter registers (offsets from F02200)
#define A1_BASE ((UINT32)0x00)
#define INTENSITYINC ((UINT32)0x70)
#define ZINC ((UINT32)0x74)
#define COLLISIONCTRL ((UINT32)0x78)
-#define PHRASEINT3 ((UINT32)0x7C)
-#define PHRASEINT2 ((UINT32)0x80)
-#define PHRASEINT1 ((UINT32)0x84)
-#define PHRASEINT0 ((UINT32)0x88)
-#define PHRASEZ3 ((UINT32)0x8C)
-#define PHRASEZ2 ((UINT32)0x90)
-#define PHRASEZ1 ((UINT32)0x94)
-#define PHRASEZ0 ((UINT32)0x98)
+#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)
// Blitter command bits
#define DCOMPEN (cmd & 0x08000000)
#define PATDSEL (cmd & 0x00010000)
-#define INTADD (cmd & 0x00020000)
+#define ADDDSEL (cmd & 0x00020000)
#define TOPBEN (cmd & 0x00004000)
#define TOPNEN (cmd & 0x00008000)
#define BKGWREN (cmd & 0x10000000)
#define XADD0 2
#define XADDINC 3
-#define XSIGNSUB_A1 (REG(A1_FLAGS)&0x80000)
-#define XSIGNSUB_A2 (REG(A2_FLAGS)&0x80000)
+#define XSIGNSUB_A1 (REG(A1_FLAGS)&0x080000)
+#define XSIGNSUB_A2 (REG(A2_FLAGS)&0x080000)
#define YSIGNSUB_A1 (REG(A1_FLAGS)&0x100000)
#define YSIGNSUB_A2 (REG(A2_FLAGS)&0x100000)
-#define YADD1_A1 (REG(A1_FLAGS)&0x40000)
-#define YADD1_A2 (REG(A2_FLAGS)&0x40000)
+#define YADD1_A1 (REG(A1_FLAGS)&0x040000)
+#define YADD1_A2 (REG(A2_FLAGS)&0x040000)
+//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 READ_PIXEL_1(a) ((jaguar_byte_read(a##_addr+PIXEL_OFFSET_1(a)) >> PIXEL_SHIFT_1(a)) & 0x01)
+#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 READ_PIXEL_2(a) ((jaguar_byte_read(a##_addr+PIXEL_OFFSET_2(a)) >> PIXEL_SHIFT_2(a)) & 0x03)
+#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 READ_PIXEL_4(a) ((jaguar_byte_read(a##_addr+PIXEL_OFFSET_4(a)) >> PIXEL_SHIFT_4(a)) & 0x0f)
+#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 READ_PIXEL_8(a) (jaguar_byte_read(a##_addr+PIXEL_OFFSET_8(a)))
+#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 READ_PIXEL_16(a) (jaguar_word_read(a##_addr+(PIXEL_OFFSET_16(a)<<1)))
+#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 READ_PIXEL_32(a) (jaguar_long_read(a##_addr+(PIXEL_OFFSET_32(a)<<2)))
+#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)))
// pixel read
#define READ_PIXEL(a,f) (\
// 16 bpp z data read
#define ZDATA_OFFSET_16(a) (PIXEL_OFFSET_16(a) + a##_zoffs * 4)
-#define READ_ZDATA_16(a) (jaguar_word_read(a##_addr+(ZDATA_OFFSET_16(a)<<1)))
+#define READ_ZDATA_16(a) (JaguarReadWord(a##_addr+(ZDATA_OFFSET_16(a)<<1), BLITTER))
+//#define READ_ZDATA_16(a) (JaguarReadWord(a##_addr+(ZDATA_OFFSET_16(a)<<1)))
// z data read
#define READ_ZDATA(a,f) (READ_ZDATA_16(a))
// 16 bpp z data write
-#define WRITE_ZDATA_16(a,d) { jaguar_word_write(a##_addr+(ZDATA_OFFSET_16(a)<<1),d); }
-//#define WRITE_ZDATA_16(a,d) { jaguar_word_write(a##_addr+(ZDATA_OFFSET_16(a)<<1),d); \
- WriteLog("16bpp z write --> "); }
+#define WRITE_ZDATA_16(a,d) { JaguarWriteWord(a##_addr+(ZDATA_OFFSET_16(a)<<1), d, BLITTER); }
+//#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);
// 1 bpp r data read
-#define READ_RDATA_1(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x>>19)&4)))>>(((UINT32)a##_x>>16)&0x1f))& 0x1 : (REG(r) & 0x1))
+#define READ_RDATA_1(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 19) & 0x04))) >> (((UINT32)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)&4)))>>(((UINT32)a##_x>>15)&0x3e))& 0x3 : (REG(r) & 0x3))
+#define READ_RDATA_2(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 18) & 0x04))) >> (((UINT32)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)&4)))>>(((UINT32)a##_x>>14)&0x28))& 0xf : (REG(r) & 0xf))
+#define READ_RDATA_4(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 17) & 0x04))) >> (((UINT32)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)&4)))>>(((UINT32)a##_x>>13)&0x18))& 0xff : (REG(r) & 0xff))
+#define READ_RDATA_8(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 16) & 0x04))) >> (((UINT32)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)&4)))>>(((UINT32)a##_x>>12)&0x10))&0xffff : (REG(r) & 0xffff))
+#define READ_RDATA_16(r,a,p) ((p) ? ((REG(r+(((UINT32)a##_x >> 15) & 0x04))) >> (((UINT32)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)&4)) : REG(r))
+#define READ_RDATA_32(r,a,p) ((p) ? REG(r+(((UINT32)a##_x >> 14) & 0x04)) : REG(r))
// register data read
#define READ_RDATA(r,a,f,p) (\
(((f>>3)&0x07) == 5) ? (READ_RDATA_32(r,a,p)) : 0)
// 1 bpp pixel write
-#define WRITE_PIXEL_1(a,d) { jaguar_byte_write(a##_addr+PIXEL_OFFSET_1(a),(jaguar_byte_read(a##_addr+PIXEL_OFFSET_1(a))&(~(0x01 << PIXEL_SHIFT_1(a))))|(d<<PIXEL_SHIFT_1(a))); }
+#define WRITE_PIXEL_1(a,d) { JaguarWriteByte(a##_addr+PIXEL_OFFSET_1(a), (JaguarReadByte(a##_addr+PIXEL_OFFSET_1(a), BLITTER)&(~(0x01 << PIXEL_SHIFT_1(a))))|(d<<PIXEL_SHIFT_1(a)), BLITTER); }
+//#define WRITE_PIXEL_1(a,d) { JaguarWriteByte(a##_addr+PIXEL_OFFSET_1(a), (JaguarReadByte(a##_addr+PIXEL_OFFSET_1(a))&(~(0x01 << PIXEL_SHIFT_1(a))))|(d<<PIXEL_SHIFT_1(a))); }
// 2 bpp pixel write
-#define WRITE_PIXEL_2(a,d) { jaguar_byte_write(a##_addr+PIXEL_OFFSET_2(a),(jaguar_byte_read(a##_addr+PIXEL_OFFSET_2(a))&(~(0x03 << PIXEL_SHIFT_2(a))))|(d<<PIXEL_SHIFT_2(a))); }
+#define WRITE_PIXEL_2(a,d) { JaguarWriteByte(a##_addr+PIXEL_OFFSET_2(a), (JaguarReadByte(a##_addr+PIXEL_OFFSET_2(a), BLITTER)&(~(0x03 << PIXEL_SHIFT_2(a))))|(d<<PIXEL_SHIFT_2(a)), BLITTER); }
+//#define WRITE_PIXEL_2(a,d) { JaguarWriteByte(a##_addr+PIXEL_OFFSET_2(a), (JaguarReadByte(a##_addr+PIXEL_OFFSET_2(a))&(~(0x03 << PIXEL_SHIFT_2(a))))|(d<<PIXEL_SHIFT_2(a))); }
// 4 bpp pixel write
-#define WRITE_PIXEL_4(a,d) { jaguar_byte_write(a##_addr+PIXEL_OFFSET_4(a),(jaguar_byte_read(a##_addr+PIXEL_OFFSET_4(a))&(~(0x0f << PIXEL_SHIFT_4(a))))|(d<<PIXEL_SHIFT_4(a))); }
+#define WRITE_PIXEL_4(a,d) { JaguarWriteByte(a##_addr+PIXEL_OFFSET_4(a), (JaguarReadByte(a##_addr+PIXEL_OFFSET_4(a), BLITTER)&(~(0x0f << PIXEL_SHIFT_4(a))))|(d<<PIXEL_SHIFT_4(a)), BLITTER); }
+//#define WRITE_PIXEL_4(a,d) { JaguarWriteByte(a##_addr+PIXEL_OFFSET_4(a), (JaguarReadByte(a##_addr+PIXEL_OFFSET_4(a))&(~(0x0f << PIXEL_SHIFT_4(a))))|(d<<PIXEL_SHIFT_4(a))); }
// 8 bpp pixel write
-#define WRITE_PIXEL_8(a,d) { jaguar_byte_write(a##_addr+PIXEL_OFFSET_8(a),d); }
+#define WRITE_PIXEL_8(a,d) { JaguarWriteByte(a##_addr+PIXEL_OFFSET_8(a), d, BLITTER); }
+//#define WRITE_PIXEL_8(a,d) { JaguarWriteByte(a##_addr+PIXEL_OFFSET_8(a), d); }
// 16 bpp pixel write
-#define WRITE_PIXEL_16(a,d) { jaguar_word_write(a##_addr+(PIXEL_OFFSET_16(a)<<1),d); }
-//#define WRITE_PIXEL_16(a,d) { jaguar_word_write(a##_addr+(PIXEL_OFFSET_16(a)<<1),d); \
- WriteLog("16bpp pixel write --> "); }
+//#define WRITE_PIXEL_16(a,d) { JaguarWriteWord(a##_addr+(PIXEL_OFFSET_16(a)<<1),d); }
+#define WRITE_PIXEL_16(a,d) { JaguarWriteWord(a##_addr+(PIXEL_OFFSET_16(a)<<1), d, BLITTER); if (specialLog) WriteLog("Pixel write address: %08X\n", a##_addr+(PIXEL_OFFSET_16(a)<<1)); }
+//#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)); }
-//This is where the bad YPOS values are being written... How to fix???
// 32 bpp pixel write
-#define WRITE_PIXEL_32(a,d) { jaguar_long_write(a##_addr+(PIXEL_OFFSET_32(a)<<2),d); }
-//#define WRITE_PIXEL_32(a,d) { jaguar_long_write(a##_addr+(PIXEL_OFFSET_32(a)<<2),d); \
- WriteLog("32bpp 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); }
// pixel write
#define WRITE_PIXEL(a,f,d) {\
case 5: WRITE_PIXEL_32(a,d); break; \
}}
+// External global variables
+
+extern int jaguar_active_memory_dumps;
+
+// Local global variables
+
+int start_logging = 0;
+uint8 blitter_working = 0;
+
+// Blitter register RAM (most of it is hidden from the user)
+
+static uint8 blitter_ram[0x100];
+
// Width in Pixels of a Scanline
// This is a pretranslation of the value found in the A1 & A2 flags: It's really a floating point value
// of the form EEEEMM where MM is the mantissa with an implied "1." in front of it and the EEEE value is
// 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 blitter_scanline_width[48] =
{
0, 0, 0, 0, // Note: This would really translate to 1, 1, 1, 1
2, 0, 0, 0,
512, 640, 768, 896,
1024, 1280, 1536, 1792,
2048, 2560, 3072, 3584
-};
+};//*/
//static uint8 * tom_ram_8;
//static uint8 * paletteRam;
//static uint16 gint[4];
//static uint16 gfrac[4];
//static uint8 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 uint32 a1_clip_x, a1_clip_y;
-
-static uint8 blitter_code_cache[4096];
-static uint8 * blitter_ptr;
-uint8 blitter_working = 0;
-
-typedef void (blitter_fn)(void);
-
-typedef struct s_blitter_cache
-{
- uint32 hashcode;
- uint8 * code;
- uint32 ready;
- uint8 used;
- struct s_blitter_cache * next;
- struct s_blitter_cache * prev;
-} s_blitter_code_cache;
-
-s_blitter_cache * blitter_cache[256];
-
-uint8 blitter_cache_init = 0;
-static uint8 BPP_LUT[8] = { 1, 2, 4, 8, 16, 32, 0, 0 };
-
-FILE * blitters_code_fp;
-FILE * blitters_code_init_fp;
-
-//////////////////////////////////////////////////////////////////////////////
-// build C code for the specified blitter
-//////////////////////////////////////////////////////////////////////////////
-//
-//
-//
-//////////////////////////////////////////////////////////////////////////////
-void blitter_gen_c_code(FILE *fp, uint32 cmd,uint32 hashcode)
-{
- static uint8 inhibit_modified=0;
-
- fprintf(fp,"#ifndef blitter_code_0x%.8x\n",hashcode);
- fprintf(fp,"#define blitter_code_0x%.8x\n",hashcode);
-
- fprintf(fp,"void blitter_0x%.8x(void)\n",hashcode);
- fprintf(fp,"{\n");
- fprintf(fp,"\twhile (outer_loop--)\n");
- fprintf(fp,"\t{\n");
- fprintf(fp,"\t\tinner_loop=n_pixels;\n");
- fprintf(fp,"\t\twhile (inner_loop--)\n");
- fprintf(fp,"\t\t{\n");
- fprintf(fp,"\t\t\tuint32 srcdata = 0;\n");
- fprintf(fp,"\t\t\tuint32 srczdata = 0;\n");
- fprintf(fp,"\t\t\tuint32 dstdata = 0;\n");
- fprintf(fp,"\t\t\tuint32 dstzdata = 0;\n");
- fprintf(fp,"\t\t\tuint32 writedata = 0;\n");
- fprintf(fp,"\t\t\tuint32 inhibit = 0;\n");
-
- char *src;
- char *dst;
- uint32 src_flags;
- uint32 dst_flags;
-
- if (!DSTA2)
- {
- src="a2";
- dst="a1";
- src_flags=A2_FLAGS;
- dst_flags=A1_FLAGS;
- }
- else
- {
- src="a1";
- dst="a2";
- src_flags=A1_FLAGS;
- dst_flags=A2_FLAGS;
- }
-
- // load src data and Z
- if (SRCEN)
- {
- fprintf(fp, "\t\t\tsrcdata = READ_PIXEL_%i(%s);\n", BPP_LUT[(((REG(dst_flags)) >> 3) & 0x07)], src);
- if (SRCENZ)
- fprintf(fp,"\t\t\tsrczdata = READ_ZDATA_%i(%s);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],src);
- else
- if (cmd & 0x001c020)
- fprintf(fp,"\t\t\tsrczdata = READ_RDATA_%i(SRCZINT, %s, %s_phrase_mode);\n",BPP_LUT[(((REG(src_flags))>>3)&0x07)],src,src);
- }
- else
- {
- fprintf(fp,"\t\t\tsrcdata = READ_RDATA_%i(SRCDATA, %s, %s_phrase_mode);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],src,src);
- if (cmd & 0x001c020)
- fprintf(fp,"\t\t\tsrczdata = READ_RDATA_%i(SRCZINT, %s, %s_phrase_mode);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],src,src);
- }
-
- // load dst data and Z
- if (DSTEN)
- {
- fprintf(fp,"\t\t\tdstdata = READ_PIXEL_%i(%s);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst);
- if (DSTENZ)
- fprintf(fp,"\t\t\tdstzdata = READ_ZDATA_%i(%s);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst);
- else
- fprintf(fp,"\t\t\tdstzdata = READ_RDATA_%i(DSTZ, %s, %s_phrase_mode);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst,dst);
- }
- else
- {
- fprintf(fp,"\t\t\tdstdata = READ_RDATA_%i(DSTDATA, %s, %s_phrase_mode);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst,dst);
-
- if (DSTENZ)
- fprintf(fp,"\t\t\tdstzdata = READ_RDATA_%i(DSTZ, %s, %s_phrase_mode);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst,dst);
- }
-
- // a1 clipping
- if ((cmd & 0x00000040)&&(!DSTA2))
-
- {
- fprintf(fp,"\t\t\tif (a1_x < 0 || a1_y < 0 || (a1_x >> 16) >= (REG(A1_CLIP) & 0x7fff) || (a1_y >> 16) >= ((REG(A1_CLIP) >> 16) & 0x7fff)) inhibit = 1;\n");
- inhibit_modified=1;
- }
- if(GOURZ)
- {
- fprintf(fp,"\t\t\tsrczdata=z_i[colour_index]>>16;\n");
- }
- // apply z comparator
- if (Z_OP_INF) { fprintf(fp,"\t\t\tif (srczdata < dstzdata) inhibit = 1;\n"); inhibit_modified=1;}
- if (Z_OP_EQU) { fprintf(fp,"\t\t\tif (srczdata == dstzdata) inhibit = 1;\n"); inhibit_modified=1;}
- if (Z_OP_SUP) { fprintf(fp,"\t\t\tif (srczdata > dstzdata) inhibit = 1;\n"); inhibit_modified=1;}
-
- // apply data comparator
- if (DCOMPEN)
- {
- if (!CMPDST)
- {
- // compare source pixel with pattern pixel
- fprintf(fp,"\t\t\tif (srcdata == READ_RDATA_%i(PATTERNDATA, %s,%s_phrase_mode)) inhibit=1;\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],src,src);
- inhibit_modified=1;
- }
- else
- {
- // compare destination pixel with pattern pixel
- fprintf(fp,"\t\t\tif (dstdata == READ_RDATA_%i(PATTERNDATA, %s,%s_phrase_mode)) inhibit=1;\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst,dst);
- inhibit_modified=1;
- }
- }
-
- // compute the write data and store
- if (inhibit_modified) fprintf(fp,"\t\t\tif (!inhibit)\n\t\t\t{\n");
- if (PATDSEL)
- {
- // use pattern data for write data
- fprintf(fp,"\t\t\t\twritedata= READ_RDATA_%i(PATTERNDATA, %s, %s_phrase_mode);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst,dst);
- }
- else
- if (INTADD)
- {
- // intensity addition
- fprintf(fp,"\t\t\t\twritedata = (srcdata & 0xff) + (dstdata & 0xff);\n");
- if (!(TOPBEN))
- fprintf(fp,"\t\t\t\tif (writedata > 0xff) writedata = 0xff;\n");
-
- fprintf(fp,"\t\t\t\twritedata |= (srcdata & 0xf00) + (dstdata & 0xf00);\n");
- if (!(TOPNEN)) fprintf(fp,"\t\t\t\tif (writedata > 0xfff) writedata = 0xfff;\n");
- fprintf(fp,"\t\t\t\twritedata |= (srcdata & 0xf000) + (dstdata & 0xf000);\n");
- }
- else
- {
- if (LFU_NAN) fprintf(fp,"\t\t\t\twritedata |= ~srcdata & ~dstdata;\n");
- if (LFU_NA) fprintf(fp,"\t\t\t\twritedata |= ~srcdata & dstdata;\n");
- if (LFU_AN) fprintf(fp,"\t\t\t\twritedata |= srcdata & ~dstdata;\n");
- if (LFU_A) fprintf(fp,"\t\t\t\twritedata |= srcdata & dstdata;\n");
- }
- if(GOURD)
- {
- fprintf(fp,"\t\t\t\twritedata = ((gd_c[colour_index])<<8)|(gd_i[colour_index]>>16);\n");
- }
- if(SRCSHADE)
- {
- fprintf(fp,"\t\t\t\t{\n");
- fprintf(fp,"\t\t\t\tint intensity = srcdata & 0xFF;\n");
- fprintf(fp,"\t\t\t\tint ia = gd_ia >> 16;\n");
- fprintf(fp,"\t\t\t\tif(ia & 0x80)\n");
- fprintf(fp,"\t\t\t\t ia = 0xFFFFFF00 | ia;\n");
- fprintf(fp,"\t\t\t\tintensity += ia;\n");
- fprintf(fp,"\t\t\t\tif(intensity < 0)\n");
- fprintf(fp,"\t\t\t\t intensity = 0;\n");
- fprintf(fp,"\t\t\t\tif(intensity > 0xFF)\n");
- fprintf(fp,"\t\t\t\t intensity = 0xFF;\n");
- fprintf(fp,"\t\t\t\twritedata = (srcdata & 0xFF00) | intensity;\n");
- fprintf(fp,"\t\t\t\t}\n");
- }
- if (inhibit_modified)
- {
- fprintf(fp,"\t\t\t} else { srczdata=dstzdata; writedata=dstdata; }\n");
- }
+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];
- if ((DSTA2?a2_phrase_mode:a1_phrase_mode) || BKGWREN)
- {
- // write to the destination
- fprintf(fp,"\t\t\tWRITE_PIXEL_%i(%s, writedata);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst);
- if (DSTWRZ) fprintf(fp,"\t\t\tWRITE_ZDATA_%i(%s, srczdata);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst);
- }
- else
- {
- if (inhibit_modified) fprintf(fp,"\t\t\tif (!inhibit)\n\t\t\t{\n");
- // write to the destination
- fprintf(fp,"\t\t\t\tWRITE_PIXEL_%i(%s, writedata);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst);
- if (DSTWRZ) fprintf(fp,"\t\t\t\tWRITE_ZDATA_%i(%s, srczdata);\n",BPP_LUT[(((REG(dst_flags))>>3)&0x07)],dst);
- if (inhibit_modified) fprintf(fp,"\t\t\t}\n");
- }
- // update x and y
- fprintf(fp,"\t\t\ta1_x += a1_xadd;\n");
- fprintf(fp,"\t\t\ta1_y += a1_yadd;\n");
- fprintf(fp,"\t\t\ta2_x = (a2_x + a2_xadd) & a2_mask_x;\n");
- fprintf(fp,"\t\t\ta2_y = (a2_y + a2_yadd) & a2_mask_y;\n");
- if (GOURZ)
- {
- fprintf(fp,"\t\t\tz_i[colour_index]+=zadd;\n");
- }
- if ((GOURD)||(SRCSHADE))
- {
- fprintf(fp,"\t\t\tgd_i[colour_index] += gd_ia;\n");
- fprintf(fp,"\t\t\tgd_c[colour_index] += gd_ca;\n");
- }
- if ((GOURD)||(SRCSHADE)||(GOURZ))
- {
- if (a1_phrase_mode)
- fprintf(fp,"\t\t\t colour_index=(colour_index+1)&0x3;\n");
- }
- fprintf(fp,"\t\t}\n");
+static int32 a1_clip_x, a1_clip_y;
- fprintf(fp,"\t\ta1_x+=a1_step_x;\n");
- fprintf(fp,"\t\ta1_y+=a1_step_y;\n");
- fprintf(fp,"\t\ta2_x+=a2_step_x;\n");
- fprintf(fp,"\t\ta2_y+=a2_step_y;\n");
- fprintf(fp,"\t}\n");
-
- // write values back to registers
- fprintf(fp,"\tWREG(A1_PIXEL, (a1_y & 0xffff0000) | ((a1_x >> 16) & 0xffff));\n");
- fprintf(fp,"\tWREG(A1_FPIXEL, (a1_y << 16) | (a1_x & 0xffff));\n");
- fprintf(fp,"\tWREG(A2_PIXEL, (a2_y & 0xffff0000) | ((a2_x >> 16) & 0xffff));\n");
- fprintf(fp,"}\n");
- fprintf(fp,"#endif\n");
-}
+// 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
+// to optimize the blitter, then we may revisit it in the future...
-//////////////////////////////////////////////////////////////////////////////
-// Generate a start of function in x86 assembly
-//////////////////////////////////////////////////////////////////////////////
-//
-//
-//
-//////////////////////////////////////////////////////////////////////////////
-void blitter_gen_start_of_function(void)
-{
- *blitter_ptr++ = 0x55; // push ebp
- *blitter_ptr++ = 0x8B; // mov ebp,esp
- *blitter_ptr++ = 0xEC;
-}
-//////////////////////////////////////////////////////////////////////////////
-// Generate a end of function in x86 assembly
-//////////////////////////////////////////////////////////////////////////////
-//
//
+// Generic blit handler
//
-//////////////////////////////////////////////////////////////////////////////
-void blitter_gen_end_of_function(void)
+void blitter_generic(uint32 cmd)
{
- *blitter_ptr++ = 0x8B; // mov esp,ebp
- *blitter_ptr++ = 0xE5;
- *blitter_ptr++ = 0x5D; // pop ebp
- *blitter_ptr++ = 0xC3; // ret
-}
-
-#define HASHCODE_BIT(C,B) if (C) hashcode|=(1<<B);
-#define HASHCODE_BIT_TEST(B) (hashcode&(1<<B))
+/*
+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
+ 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
+ A2 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 192 (1E), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 100/12, A2 x/y: 106/0 Pattern: 000000F300000000
+*/
+//if (effect_start)
+// specialLog = true;
+/*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 blitter_calc_hashcode(uint32 cmd)
+if (specialLog)
{
- uint32 hashcode=0x00000000;
-
- // source and destination bit depth
- hashcode|=((REG(A1_FLAGS)>>3)&0x07)<<0;
- hashcode|=((REG(A2_FLAGS)>>3)&0x07)<<3;
-
- HASHCODE_BIT(DSTA2, 6);
- HASHCODE_BIT(SRCEN, 7);
- HASHCODE_BIT(SRCENZ, 8);
- HASHCODE_BIT(DSTEN, 9);
- HASHCODE_BIT(DSTENZ, 10);
- HASHCODE_BIT(Z_OP_INF, 11);
- HASHCODE_BIT(Z_OP_EQU, 12);
- HASHCODE_BIT(Z_OP_SUP, 13);
- HASHCODE_BIT(DCOMPEN, 14);
- HASHCODE_BIT(CMPDST, 15);
- HASHCODE_BIT(PATDSEL, 16);
- HASHCODE_BIT(INTADD, 17);
- HASHCODE_BIT(TOPBEN, 18);
- HASHCODE_BIT(TOPNEN, 19);
- HASHCODE_BIT(LFU_NAN, 20);
- HASHCODE_BIT(LFU_NA, 21);
- HASHCODE_BIT(LFU_AN, 22);
- HASHCODE_BIT(LFU_A, 23);
- HASHCODE_BIT(BKGWREN, 24);
- HASHCODE_BIT(DSTWRZ, 25);
- HASHCODE_BIT((cmd & 0x001c020), 26); // extra data read/write
- HASHCODE_BIT((cmd & 0x00000040), 27); // source clipping
- HASHCODE_BIT(a1_phrase_mode, 28);
- HASHCODE_BIT(a2_phrase_mode, 29);
-
- return(hashcode);
+ WriteLog("About to do n x m blit (BM width is ? pixels)...\n");
+ WriteLog("A1_STEP_X/Y = %08X/%08X, A2_STEP_X/Y = %08X/%08X\n", a1_step_x, a1_step_y, a2_step_x, a2_step_y);
}
-//////////////////////////////////////////////////////////////////////////////
-// Build the blitter code for the current blitter operation in the cache
-//////////////////////////////////////////////////////////////////////////////
-//
-//
-//
-//////////////////////////////////////////////////////////////////////////////
-void blitter_build_cached_code(uint32 cmd, uint32 cache_index)
-{
-}
-//////////////////////////////////////////////////////////////////////////////
-// Check if the blitter code for the current blitter operation is cached
-//////////////////////////////////////////////////////////////////////////////
-//
-//
-//
-//////////////////////////////////////////////////////////////////////////////
-struct s_blitter_cache * blitter_in_cache(uint32 cmd)
-{
- uint32 i;
- uint32 hashcode=blitter_calc_hashcode(cmd);
-#ifdef LOG_BLITS
- WriteLog("blitter: hashcode= 0x%.8x\n",hashcode);
-#endif
- struct s_blitter_cache *blitter_list=blitter_cache[hashcode>>24];
-
- i=0;
- while (blitter_list->next)
+/* if (BCOMPEN)
{
- blitter_list=blitter_list->next;
-
- if (blitter_list->hashcode==hashcode)
- return(blitter_list);
- }
-#ifdef GEN_CODE
- blitter_list->next=(struct s_blitter_cache *)malloc(sizeof(struct s_blitter_cache));
- blitter_list->next->prev=blitter_list;
- blitter_list->next->next=null;
- blitter_list=blitter_list->next;
-
- blitter_list->code=(uint8*)malloc(4096);
- blitter_list->hashcode=hashcode;
- blitter_list->ready=0;
- blitter_gen_c_code(blitters_code_fp,cmd,hashcode);
- fprintf(blitters_code_init_fp,"\tblitter_add(0x%.8x,(uint8*)&blitter_0x%.8x);\n",hashcode,hashcode);
-#else
- //WriteLog("warning: using generic blitter for blitter 0x%.8x\n",hashcode);
-#endif
- return(null);
-}
-#ifndef USE_GENERIC_BLITTER
-#include "include/blit_c.h"
-#endif
-//////////////////////////////////////////////////////////////////////////////
-// Execute the cached blitter code for the current blitter operation
-//////////////////////////////////////////////////////////////////////////////
-//
-//
-//
-//////////////////////////////////////////////////////////////////////////////
-uint32 blitter_execute_cached_code(struct s_blitter_cache * blitter)
-{
- if ((blitter == null) || (blitter->ready == 0))
- return 0;
-
- blitter_fn * fn = (blitter_fn *)blitter->code;
- blitter->used = 1;
- (*fn)();
-
- return 1;
-}
-
-void blitter_add(uint32 hashcode, uint8 *code)
-{
- struct s_blitter_cache *blitter_list=blitter_cache[(hashcode>>24)];
-
-// WriteLog("adding blitter for hashcode 0x%.8x\n",hashcode);
+ if (DSTA2)
+ a1_xadd = 0;
+ else
+ a2_xadd = 0;
+ }//*/
- while (blitter_list->next)
+ while (outer_loop--)
{
- blitter_list=blitter_list->next;
-
- if (blitter_list->hashcode==hashcode)
- return;
- }
- blitter_list->next=(struct s_blitter_cache *)malloc(sizeof(struct s_blitter_cache));
- blitter_list->next->prev=blitter_list;
- blitter_list->next->next=null;
- blitter_list=blitter_list->next;
-
- blitter_list->code=code;
- blitter_list->hashcode=hashcode;
- blitter_list->ready=1;
- blitter_list->used=0;
-}
-
-void blitter_list(void)
+if (specialLog)
{
-/*
- WriteLog("Used blitters list:\n");
-
- for (int i=0;i<256;i++)
- {
- struct s_blitter_cache *blitter_list=blitter_cache[i];
+ 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;
- while (blitter_list->next)
+ //Kludge for Hover Strike...
+ //I wonder if this kludge is in conjunction with the SRCENX down below...
+ // This isn't so much a kludge but the way things work in BCOMPEN mode...!
+ if (BCOMPEN && SRCENX)
{
- blitter_list=blitter_list->next;
- if (blitter_list->used)
- WriteLog("\t0%.8x\n",blitter_list->hashcode);
+ if (n_pixels < bppSrc)
+ bitPos = bppSrc - n_pixels;
}
- }
-*/
-}
-
-//
-// Generic blit handler
-//
-void blitter_generic(uint32 cmd)
-{
-//Testing only!
-//uint32 logGo = ((cmd == 0x01800E01 && REG(A1_BASE) == 0x898000) ? 1 : 0);
-
-/* uint32 srcdata = 0;
- uint32 srczdata = 0;
- uint32 dstdata = 0;
- uint32 dstzdata = 0;
- uint32 writedata = 0;
- uint32 inhibit = 0;*/
- uint32 srcdata, srczdata, dstdata, dstzdata, writedata, inhibit;
-
- while (outer_loop--)
- {
inner_loop = n_pixels;
while (inner_loop--)
{
+if (specialLog)
+{
+ WriteLog(" A1_X/Y = %08X/%08X, A2_X/Y = %08X/%08X\n", a1_x, a1_y, a2_x, a2_y);
+}
srcdata = srczdata = dstdata = dstzdata = writedata = inhibit = 0;
- if (!DSTA2)
+ if (!DSTA2) // Data movement: A1 <- A2
{
// load src data and Z
- if (SRCEN)
+// if (SRCEN)
+ if (SRCEN || SRCENX) // Not sure if this is correct... (seems to be...!)
{
srcdata = READ_PIXEL(a2, REG(A2_FLAGS));
if (SRCENZ)
srczdata = READ_ZDATA(a2, REG(A2_FLAGS));
- else if (cmd & 0x0001C020)
+ else if (cmd & 0x0001C020) // PATDSEL | TOPBEN | TOPNEN | DSTWRZ
srczdata = READ_RDATA(SRCZINT, a2, REG(A2_FLAGS), a2_phrase_mode);
}
- else
+ else // Use SRCDATA register...
{
srcdata = READ_RDATA(SRCDATA, a2, REG(A2_FLAGS), a2_phrase_mode);
- if (cmd & 0x0001C020)
+ if (cmd & 0x0001C020) // PATDSEL | TOPBEN | TOPNEN | DSTWRZ
srczdata = READ_RDATA(SRCZINT, a2, REG(A2_FLAGS), a2_phrase_mode);
}
// apply data comparator
// Note: DCOMPEN only works in 8/16 bpp modes! !!! FIX !!!
// Does BCOMPEN only work in 1 bpp mode???
+// No, but it always does a 1 bit expansion no matter what the BPP of the channel is set to. !!! FIX !!!
+// This is bit tricky... We need to fix the XADD value so that it acts like a 1BPP value while inside
+// an 8BPP space.
if (DCOMPEN | BCOMPEN)
{
+//Temp, for testing Hover Strike
+//Doesn't seem to do it... Why?
+//What needs to happen here is twofold. First, the address generator in the outer loop has
+//to honor the BPP when calculating the start address (which it kinda does already). Second,
+//it has to step bit by bit when using BCOMPEN. How to do this???
+ if (BCOMPEN)
+//small problem with this approach: it's not accurate... We need a proper address to begin with
+//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);
+ srcdata = (srcdata >> pixShift) & 0x01;
+
+ bitPos++;
+// if (bitPos % bppSrc == 0)
+// a2_x += 0x00010000;
+ }
+/*
+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
+ 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
+ A2 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 192 (1E), 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 192 (1E), 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 192 (1E), 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 192 (1E), 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 192 (1E), 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 8bpp, z-off: 0, width: 192 (1E), addctl: XADDPIX YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 128/12, A2 x/y: 67/0 Pattern: 000000F300000000
+*/
+
+
if (!CMPDST)
{
//WriteLog("Blitter: BCOMPEN set on command %08X inhibit prev:%u, now:", cmd, inhibit);
else
WriteLog("Blitter: Bad BPP (%u) selected for BCOMPEN mode!\n", A2bpp);//*/
// What it boils down to is this:
-// ***CHECK*** Hmm. Seems to cause Rayman to freeze. Investigate.
-// This doesn't seem to be it. Hmm.
-// It was a bug in the TOM read word code (reading VC)
+
if (srcdata == 0)
inhibit = 1;//*/
}
// if (dstdata != READ_RDATA(PATTERNDATA, a1, REG(A1_FLAGS), a1_phrase_mode))
inhibit = 1;
}
+
// This is DEFINITELY WRONG
- if (a1_phrase_mode || a2_phrase_mode)
- inhibit = !inhibit;
+// if (a1_phrase_mode || a2_phrase_mode)
+// inhibit = !inhibit;
}
if (CLIPA1)
// 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:
+// 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
+
if (PATDSEL)
{
// use pattern data for write data
writedata = READ_RDATA(PATTERNDATA, a1, REG(A1_FLAGS), a1_phrase_mode);
}
- else if (INTADD)
+ else if (ADDDSEL)
{
+/*if (blit_start_log)
+ WriteLog("BLIT: ADDDSEL srcdata: %08X\, dstdata: %08X, ", srcdata, dstdata);//*/
+
// intensity addition
- writedata = (srcdata & 0xFF) + (dstdata & 0xFF);
+//Ok, this is wrong... Or is it? Yes, it's wrong! !!! FIX !!!
+/* writedata = (srcdata & 0xFF) + (dstdata & 0xFF);
if (!(TOPBEN) && writedata > 0xFF)
- writedata = 0xFF;
+// writedata = 0xFF;
+ writedata &= 0xFF;
writedata |= (srcdata & 0xF00) + (dstdata & 0xF00);
if (!(TOPNEN) && writedata > 0xFFF)
- writedata = 0xFFF;
+// writedata = 0xFFF;
+ writedata &= 0xFFF;
+ writedata |= (srcdata & 0xF000) + (dstdata & 0xF000);//*/
+//notneeded--writedata &= 0xFFFF;
+/*if (blit_start_log)
+ WriteLog("writedata: %08X\n", writedata);//*/
+/*
+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:
+ 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
+ A1 x/y: 0/0, A2 x/y: 3288/0 Pattern: 0000000000000000 SRCDATA: 00FD00FD00FD00FD
+*/
+ writedata = (srcdata & 0xFF) + (dstdata & 0xFF);
+
+ if (!TOPBEN)
+ {
+//This is correct now, but slow...
+ int16 s = (srcdata & 0xFF) | (srcdata & 0x80 ? 0xFF00 : 0x0000),
+ d = dstdata & 0xFF;
+ int16 sum = s + d;
+
+ if (sum < 0)
+ writedata = 0x00;
+ else if (sum > 0xFF)
+ writedata = 0xFF;
+ else
+ writedata = (uint32)sum;
+ }
+
+//This doesn't seem right... Looks like it would muck up the low byte... !!! FIX !!!
+ writedata |= (srcdata & 0xF00) + (dstdata & 0xF00);
+
+ if (!TOPNEN && writedata > 0xFFF)
+ {
+ writedata &= 0xFFF;
+ }
+
writedata |= (srcdata & 0xF000) + (dstdata & 0xF000);
}
else
if (LFU_A) writedata |= srcdata & dstdata;
}
+//Although, this looks like it's OK... (even if it is shitty!)
+//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)
writedata = ((gd_c[colour_index]) << 8) | (gd_i[colour_index] >> 16);
srczdata = dstzdata;
}
+//Tried 2nd below for Hover Strike: No dice.
if (/*a1_phrase_mode || */BKGWREN || !inhibit)
+// if (/*a1_phrase_mode || BKGWREN ||*/ !inhibit)
{
-// This is the sole source of the bogus YPOS values being written to the object list... !!! FIX !!!
/*if (((REG(A1_FLAGS) >> 3) & 0x07) == 5)
{
uint32 offset = a1_addr+(PIXEL_OFFSET_32(a1)<<2);
WRITE_ZDATA(a1, REG(A1_FLAGS), srczdata);
}
}
- else // if (DSTA2)
+ else // if (DSTA2) // Data movement: A1 -> A2
{
// load src data and Z
if (SRCEN)
srcdata = READ_PIXEL(a1, REG(A1_FLAGS));
if (SRCENZ)
srczdata = READ_ZDATA(a1, REG(A1_FLAGS));
- else if (cmd & 0x0001C020)
+ else if (cmd & 0x0001C020) // PATDSEL | TOPBEN | TOPNEN | DSTWRZ
srczdata = READ_RDATA(SRCZINT, a1, REG(A1_FLAGS), a1_phrase_mode);
}
else
{
srcdata = READ_RDATA(SRCDATA, a1, REG(A1_FLAGS), a1_phrase_mode);
- if (cmd & 0x001C020)
+ if (cmd & 0x001C020) // PATDSEL | TOPBEN | TOPNEN | DSTWRZ
srczdata = READ_RDATA(SRCZINT, a1, REG(A1_FLAGS), a1_phrase_mode);
}
else
WriteLog("Blitter: Bad BPP (%u) selected for BCOMPEN mode!\n", A1bpp);//*/
// What it boils down to is this:
-// ***CHECK*** Hmm. Seems to cause Rayman to freeze. Investigate.
-// This doesn't seem to be it. Hmm.
-// It was a bug in the TOM read word code (reading VC)
if (srcdata == 0)
inhibit = 1;//*/
}
// if (dstdata != READ_RDATA(PATTERNDATA, a2, REG(A2_FLAGS), a2_phrase_mode))
inhibit = 1;
}
- if (a1_phrase_mode || a2_phrase_mode)
- inhibit =! inhibit;
+
+// This is DEFINITELY WRONG
+// if (a1_phrase_mode || a2_phrase_mode)
+// inhibit = !inhibit;
}
if (CLIPA1)
// use pattern data for write data
writedata= READ_RDATA(PATTERNDATA, a2, REG(A2_FLAGS), a2_phrase_mode);
}
- else if (INTADD)
+ else if (ADDDSEL)
{
// intensity addition
writedata = (srcdata & 0xFF) + (dstdata & 0xFF);
}//*/
// write to the destination
WRITE_PIXEL(a2, REG(A2_FLAGS), writedata);
+
if (DSTWRZ)
WRITE_ZDATA(a2, REG(A2_FLAGS), srczdata);
}
}
- // update x and y
- a1_x += a1_xadd;
- a1_y += a1_yadd;
- a2_x = (a2_x + a2_xadd) & a2_mask_x;
- a2_y = (a2_y + a2_yadd) & a2_mask_y;
+ // Update x and y (inner loop)
+//Now it does! But crappy, crappy, crappy! !!! FIX !!! [DONE]
+//This is less than ideal, but it works...
+ if (!BCOMPEN)
+ {//*/
+ a1_x += a1_xadd, a1_y += a1_yadd;
+ a2_x = (a2_x + a2_xadd) & a2_mask_x, a2_y = (a2_y + a2_yadd) & a2_mask_y;
+ }
+ else
+ {
+ a1_y += a1_yadd, a2_y = (a2_y + a2_yadd) & a2_mask_y;
+ if (!DSTA2)
+ {
+ a1_x += a1_xadd;
+ if (bitPos % bppSrc == 0)
+ a2_x = (a2_x + a2_xadd) & a2_mask_x;
+ }
+ else
+ {
+ a2_x = (a2_x + a2_xadd) & a2_mask_x;
+ if (bitPos % bppSrc == 0)
+ a1_x += a1_xadd;
+ }
+ }//*/
if (GOURZ)
z_i[colour_index] += zadd;
if (GOURD || SRCSHADE)
{
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)
+ 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)
+ gd_c[colour_index] = 0;
+if (gd_c[colour_index] > 0x000000FF)
+ gd_c[colour_index] = 0x000000FF;//*/
}
+
if (GOURD || SRCSHADE || GOURZ)
{
if (a1_phrase_mode)
+//This screws things up WORSE (for the BIOS opening screen)
+// if (a1_phrase_mode || a2_phrase_mode)
colour_index = (colour_index + 1) & 0x03;
}
}
+/*
+Here's the problem... The phrase mode code!
+Blit! (00100000 -> 00148000) count: 327 x 267, A1/2_FLAGS: 00004420/00004420 [cmd: 41802E01]
+ CMD -> src: SRCEN dst: misc: a1ctl: UPDA1 UPDA2 mode: DSTA2 GOURZ ity: z-op: op: LFU_REPLACE ctrl: SRCSHADE
+ A1 step values: -327 (X), 1 (Y)
+ A2 step values: -327 (X), 1 (Y) [mask (unused): 00000000 - FFFFFFFF/FFFFFFFF]
+ A1 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 384 (22), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 384 (22), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 28/58, A2 x/y: 28/58 Pattern: 00EA7BEA77EA77EA SRCDATA: 7BFF7BFF7BFF7BFF
+
+Below fixes it, but then borks:
+; 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 173/144, A2 x/y: 4052/0
+
+Lesse, with pre-add we'd have:
+
+ oooooooooooo
+00001111222233334444555566667777
+ ^ ^starts here...
+ | ^ends here.
+ |rolls back to here. Hmm.
+
+*/
a1_x += a1_step_x;
a1_y += a1_step_y;
a2_x += a2_step_x;
- a2_y += a2_step_y;
+ a2_y += a2_step_y;//*/
-/* if (a2_phrase_mode)
+ //New: Phrase mode taken into account! :-p
+/* if (a1_phrase_mode) // v1
{
- a1_x+=(64/a1_psize)*a1_xadd;
- }
- if (a2_phrase_mode)
+ // Bump the pointer to the next phrase boundary
+ // Even though it works, this is crappy... Clean it up!
+ uint32 size = 64 / a1_psize;
+
+ // Crappy kludge... ('aligning' source to destination)
+ if (a2_phrase_mode && DSTA2)
+ {
+ uint32 extra = (a2_start >> 16) % size;
+ a1_x += extra << 16;
+ }
+
+ uint32 newx = (a1_x >> 16) / size;
+ uint32 newxrem = (a1_x >> 16) % size;
+ a1_x &= 0x0000FFFF;
+ a1_x |= (((newx + (newxrem == 0 ? 0 : 1)) * size) & 0xFFFF) << 16;
+ }//*/
+ if (a1_phrase_mode) // v2
{
- for (int nb=0;nb<(64/a2_psize)+1;nb++)
- a2_x = (a2_x + a2_xadd) & a2_mask_x;
+ // Bump the pointer to the next phrase boundary
+ // Even though it works, this is crappy... Clean it up!
+ uint32 size = 64 / a1_psize;
+
+ // Crappy kludge... ('aligning' source to destination)
+ if (a2_phrase_mode && DSTA2)
+ {
+ uint32 extra = (a2_start >> 16) % size;
+ a1_x += extra << 16;
+ }
+
+ uint32 pixelSize = (size - 1) << 16;
+ a1_x = (a1_x + pixelSize) & ~pixelSize;
}
-*/ }
+
+/* if (a2_phrase_mode) // v1
+ {
+ // Bump the pointer to the next phrase boundary
+ // Even though it works, this is crappy... Clean it up!
+ uint32 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;
+ a2_x += extra << 16;
+ }
+
+ uint32 newx = (a2_x >> 16) / size;
+ uint32 newxrem = (a2_x >> 16) % size;
+ a2_x &= 0x0000FFFF;
+ a2_x |= (((newx + (newxrem == 0 ? 0 : 1)) * size) & 0xFFFF) << 16;
+ }//*/
+ if (a2_phrase_mode) // v1
+ {
+ // Bump the pointer to the next phrase boundary
+ // Even though it works, this is crappy... Clean it up!
+ uint32 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;
+ a2_x += extra << 16;
+ }
+
+ uint32 pixelSize = (size - 1) << 16;
+ a2_x = (a2_x + pixelSize) & ~pixelSize;
+ }
+
+ //Not entirely: This still mucks things up... !!! FIX !!!
+ //Should this go before or after the phrase mode mucking around?
+/* a1_x += a1_step_x;
+ a1_y += a1_step_y;
+ a2_x += a2_step_x;
+ a2_y += a2_step_y;//*/
+ }
// 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)
{
+//Apparently this is doing *something*, just not sure exactly what...
+/*if (cmd == 0x41802E01)
+{
+ WriteLog("BLIT: Found our blit. Was: %08X ", cmd);
+ cmd = 0x01800E01;
+ WriteLog("Is: %08X\n", cmd);
+}//*/
+
+ uint32 pitchValue[4] = { 0, 1, 3, 2 };
colour_index = 0;
src = cmd & 0x07;
dst = (cmd >> 3) & 0x07;
op = (cmd >> 21) & 0x0F;
ctrl = (cmd >> 25) & 0x3F;
- a1_addr = REG(A1_BASE);
- a2_addr = REG(A2_BASE);
+ // Addresses in A1/2_BASE are *phrase* aligned, i.e., bottom three bits are ignored!
+ // NOTE: This fixes Rayman's bad collision detection AND keeps T2K working!
+ a1_addr = REG(A1_BASE) & 0xFFFFFFF8;
+ a2_addr = REG(A2_BASE) & 0xFFFFFFF8;
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_pitch = (REG(A1_FLAGS) & 3) ^ ((REG(A1_FLAGS) & 2) >> 1);
- a2_pitch = (REG(A2_FLAGS) & 3) ^ ((REG(A2_FLAGS) & 2) >> 1);
+// a1_pitch = (REG(A1_FLAGS) & 3) ^ ((REG(A1_FLAGS) & 2) >> 1);
+// a2_pitch = (REG(A2_FLAGS) & 3) ^ ((REG(A2_FLAGS) & 2) >> 1);
+ a1_pitch = pitchValue[(REG(A1_FLAGS) & 0x03)];
+ a2_pitch = pitchValue[(REG(A2_FLAGS) & 0x03)];
n_pixels = REG(PIXLINECOUNTER) & 0xFFFF;
n_lines = (REG(PIXLINECOUNTER) >> 16) & 0xFFFF;
a1_x = (REG(A1_PIXEL) << 16) | (REG(A1_FPIXEL) & 0xFFFF);
a1_y = (REG(A1_PIXEL) & 0xFFFF0000) | (REG(A1_FPIXEL) >> 16);
+//According to the JTRM, X is restricted to 15 bits and Y is restricted to 12.
+//But it seems to fuck up T2K! !!! FIX !!!
+//Could it be sign extended??? Doesn't seem to be so according to JTRM
+// a1_x &= 0x7FFFFFFF, a1_y &= 0x0FFFFFFF;
+//Actually, it says that the X is 16 bits. But it still seems to mess with the Y when restricted to 12...
+// a1_y &= 0x0FFFFFFF;
+
// 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;
a1_width = ((0x04 | m) << e) >> 2;//*/
a2_x = (REG(A2_PIXEL) & 0x0000FFFF) << 16;
a2_y = (REG(A2_PIXEL) & 0xFFFF0000);
+//According to the JTRM, X is restricted to 15 bits and Y is restricted to 12.
+//But it seems to fuck up T2K! !!! FIX !!!
+// a2_x &= 0x7FFFFFFF, a2_y &= 0x0FFFFFFF;
+//Actually, it says that the X is 16 bits. But it still seems to mess with the Y when restricted to 12...
+// a2_y &= 0x0FFFFFFF;
+
// a2_width = blitter_scanline_width[((REG(A2_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 !!!
m = (REG(A2_FLAGS) >> 9) & 0x03, e = (REG(A2_FLAGS) >> 11) & 0x0F;
a2_width = ((0x04 | m) << e) >> 2;//*/
a2_mask_x = ((REG(A2_MASK) & 0x0000FFFF) << 16) | 0xFFFF;
break;
case XADDINC:
// add the contents of the increment register
- a1_xadd = (REG(A1_INC) << 16) | (REG(A1_FINC) & 0xFFFF);
+ a1_xadd = (REG(A1_INC) << 16) | (REG(A1_FINC) & 0x0000FFFF);
a1_yadd = (REG(A1_INC) & 0xFFFF0000) | (REG(A1_FINC) >> 16);
break;
}
+
+
+//Blit! (0011D000 -> 000B9600) count: 228 x 1, A1/2_FLAGS: 00073820/00064220 [cmd: 41802801]
+// 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);
+//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
+//This is it... The problem...
+//if ((a1_x >> 16) == 14368) // 14368 = $3820
+// return; //Lesse what we got...
+
if (XSIGNSUB_A1)
a1_xadd = -a1_xadd;
break;
//This really isn't a valid bit combo for A2... Shouldn't this cause the blitter to just say no?
case XADDINC:
+WriteLog("BLIT: Asked to use invalid bit combo (XADDINC) for A2...\n");
// add the contents of the increment register
// since there is no register for a2 we just add 1
- a2_xadd = 1 << 16;
+//Let's do nothing, since it's not listed as a valid bit combo...
+// a2_xadd = 1 << 16;
break;
}
+
if (XSIGNSUB_A2)
a2_xadd = -a2_xadd;
a1_clip_x = REG(A1_CLIP) & 0x7FFF,
a1_clip_y = (REG(A1_CLIP) >> 16) & 0x7FFF;
-// This phrase sizing is incorrect as well... !!! FIX !!!
+// This phrase sizing is incorrect as well... !!! FIX !!! [NOTHING TO FIX]
// Err, this is pixel size... (and it's OK)
a2_psize = 1 << ((REG(A2_FLAGS) >> 3) & 0x07);
a1_psize = 1 << ((REG(A1_FLAGS) >> 3) & 0x07);
- // zbuffering
+ // Z-buffering
if (GOURZ)
{
- zadd = jaguar_long_read(0xF02274);
+ zadd = REG(ZINC);
- for(int v=0; v<4; v++)
- z_i[v] = (int32)jaguar_long_read(0xF0228C + (v << 2));
+ for(int v=0; v<4; v++)
+ z_i[v] = REG(PHRASEZ0 + v*4);
}
+
+ // Gouraud shading
if (GOURD || GOURZ || SRCSHADE)
{
- // gouraud shading
- gouraud_add = jaguar_long_read(0xF02270);
+ gd_c[0] = blitter_ram[PATTERNDATA + 0];
+ gd_i[0] = ((uint32)blitter_ram[PATTERNDATA + 1] << 16)
+ | ((uint32)blitter_ram[SRCDATA + 0] << 8) | blitter_ram[SRCDATA + 1];
+
+ gd_c[1] = blitter_ram[PATTERNDATA + 2];
+ gd_i[1] = ((uint32)blitter_ram[PATTERNDATA + 3] << 16)
+ | ((uint32)blitter_ram[SRCDATA + 2] << 8) | blitter_ram[SRCDATA + 3];
+
+ gd_c[2] = blitter_ram[PATTERNDATA + 4];
+ gd_i[2] = ((uint32)blitter_ram[PATTERNDATA + 5] << 16)
+ | ((uint32)blitter_ram[SRCDATA + 4] << 8) | blitter_ram[SRCDATA + 5];
+
+ gd_c[3] = blitter_ram[PATTERNDATA + 6];
+ gd_i[3] = ((uint32)blitter_ram[PATTERNDATA + 7] << 16)
+ | ((uint32)blitter_ram[SRCDATA + 6] << 8) | blitter_ram[SRCDATA + 7];
+
+ gouraud_add = REG(INTENSITYINC);
- gd_c[0] = jaguar_byte_read(0xF02268);
- gd_i[0] = jaguar_byte_read(0xF02269);
- gd_i[0] <<= 16;
- gd_i[0] |= jaguar_word_read(0xF02240);
-
- gd_c[1] = jaguar_byte_read(0xF0226A);
- gd_i[1] = jaguar_byte_read(0xF0226B);
- gd_i[1] <<= 16;
- gd_i[1] |= jaguar_word_read(0xF02242);
-
- gd_c[2] = jaguar_byte_read(0xF0226C);
- gd_i[2] = jaguar_byte_read(0xF0226D);
- gd_i[2] <<= 16;
- gd_i[2] |= jaguar_word_read(0xF02244);
-
- gd_c[3] = jaguar_byte_read(0xF0226E);
- gd_i[3] = jaguar_byte_read(0xF0226F);
- gd_i[3] <<= 16;
- gd_i[3] |= jaguar_word_read(0xF02246);
-
- gd_ia = gouraud_add & 0xFFFFFF;
- if (gd_ia & 0x800000)
+ gd_ia = gouraud_add & 0x00FFFFFF;
+ if (gd_ia & 0x00800000)
gd_ia = 0xFF000000 | gd_ia;
- gd_ca = (gouraud_add>>24) & 0xFF;
- if (gd_ca & 0x80)
+ gd_ca = (gouraud_add >> 24) & 0xFF;
+ if (gd_ca & 0x00000080)
gd_ca = 0xFFFFFF00 | gd_ca;
}
- // fix for zoop! and syndicate
-/* if ((jaguar_mainRom_crc32==0x501be17c)||
- (jaguar_mainRom_crc32==0x70895c51)||
- (jaguar_mainRom_crc32==0x0f1f1497)||
- (jaguar_mainRom_crc32==0xfc8f0dcd)
- )
- {
- if (a1_step_x < 0)
- a1_step_x = (-n_pixels) * 65536;
-
- if (a2_step_x < 0)
- a2_step_x = (-n_pixels) * 65536;;
- }
- else
- // fix for wolfenstein 3d
- if (jaguar_mainRom_crc32==0x3966698f)
+ // Bit comparitor fixing...
+/* if (BCOMPEN)
{
- if (n_pixels==24)
- {
- if ((a1_step_x / 65536)==-28)
- {
- a1_step_x=-24*65536; // au lieu de -28
- a2_step_x= 0*65536; // au lieu de -8
- }
- }
- }
- else
- // fix for Tempest 2000
- if (jaguar_mainRom_crc32==0x32816d44)
+ // Determine the data flow direction...
+ if (!DSTA2)
+ a2_step_x /= (1 << ((REG(A2_FLAGS) >> 3) & 0x07));
+ else
+ ;//add this later
+ }//*/
+/* if (BCOMPEN)//Kludge for Hover Strike... !!! FIX !!!
{
-
- if ((n_lines!=1)&&((n_pixels==288)||(n_pixels==384)))
- {
- WriteLog("Blit!\n");
- WriteLog(" cmd = 0x%.8x\n",cmd);
- WriteLog(" a1_base = %08X\n", a1_addr);
- WriteLog(" a1_pitch = %d\n", a1_pitch);
- WriteLog(" a1_psize = %d\n", a1_psize);
- WriteLog(" a1_width = %d\n", a1_width);
- WriteLog(" a1_xadd = %f (phrase=%d)\n", (float)a1_xadd / 65536.0, a1_phrase_mode);
- WriteLog(" a1_yadd = %f\n", (float)a1_yadd / 65536.0);
- WriteLog(" a1_xstep = %f\n", (float)a1_step_x / 65536.0);
- WriteLog(" a1_ystep = %f\n", (float)a1_step_y / 65536.0);
- WriteLog(" a1_x = %f\n", (float)a1_x / 65536.0);
- WriteLog(" a1_y = %f\n", (float)a1_y / 65536.0);
- WriteLog(" a1_zoffs = %i\n",a1_zoffs);
-
- WriteLog(" a2_base = %08X\n", a2_addr);
- WriteLog(" a2_pitch = %d\n", a2_pitch);
- WriteLog(" a2_psize = %d\n", a2_psize);
- WriteLog(" a2_width = %d\n", a2_width);
- WriteLog(" a2_xadd = %f (phrase=%d)\n", (float)a2_xadd / 65536.0, a2_phrase_mode);
- WriteLog(" a2_yadd = %f\n", (float)a2_yadd / 65536.0);
- WriteLog(" a2_xstep = %f\n", (float)a2_step_x / 65536.0);
- WriteLog(" a2_ystep = %f\n", (float)a2_step_y / 65536.0);
- WriteLog(" a2_x = %f\n", (float)a2_x / 65536.0);
- WriteLog(" a2_y = %f\n", (float)a2_y / 65536.0);
- WriteLog(" a2_mask_x= 0x%.4x\n",a2_mask_x);
- WriteLog(" a2_mask_y= 0x%.4x\n",a2_mask_y);
- WriteLog(" a2_zoffs = %i\n",a2_zoffs);
-
- WriteLog(" count = %d x %d\n", n_pixels, n_lines);
-
- WriteLog(" command = %08X\n", cmd);
- WriteLog(" dsten = %i\n",DSTEN);
- WriteLog(" srcen = %i\n",SRCEN);
- WriteLog(" patdsel = %i\n",PATDSEL);
- WriteLog(" color = 0x%.8x\n",REG(PATTERNDATA));
- WriteLog(" dcompen = %i\n",DCOMPEN);
- WriteLog(" bcompen = %i\n",BCOMPEN);
- WriteLog(" cmpdst = %i\n",CMPDST);
- WriteLog(" GOURZ = %i\n",GOURZ);
- WriteLog(" GOURD = %i\n",GOURD);
- WriteLog(" SRCSHADE = %i\n",SRCSHADE);
- WriteLog(" DSTDATA = 0x%.8x%.8x\n",REG(DSTDATA),REG(DSTDATA+4));
- }
+ // Determine the data flow direction...
+ if (!DSTA2)
+ a2_x <<= 3;
}//*/
#ifdef LOG_BLITS
}
#endif
-extern int blit_start_log;
-extern int op_start_log;
+//NOTE: Pitch is ignored!
+
+//This *might* be the altimeter blits (they are)...
+//On captured screen, x-pos for black (inner) is 259, for pink is 257
+//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:
+ 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:
+ 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:
+ 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:
+ 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
+ A1 x/y: 264/128, A2 x/y: 336/0
+
+ 264v vCursor ends up here...
+ xxxxx...`
+ 111122223333
+
+262v vCursor ends up here...
+ xxxxxxxxx.'
+ 1111222233334444
+
+Fixed! Now for more:
+
+; 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 20/24, A2 x/y: 5780/0
+
+Also fixed!
+
+More (not sure this is a blitter problem as much as it's a GPU problem):
+All but the "M" are trashed...
+This does *NOT* look like a blitter problem, as it's rendering properly...
+Actually, if you look at the A1 step values, there IS a discrepancy!
+
+; 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 134/144, A2 x/y: 2516/0
+;129,146: +5,-2
+
+; 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 147/144, A2 x/y: 2660/0
+
+; 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 160/144, A2 x/y: 3764/0
+
+; 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
+ 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
+ A2 -> pitch: 1 phrases, depth: 16bpp, z-off: 0, width: 1 (00), addctl: XADDPHR YADD0 XSIGNADD YSIGNADD
+ A1 x/y: 173/144, A2 x/y: 4052/0
+
+*/
+//extern int op_start_log;
if (blit_start_log)
{
char * ctrlStr[4] = { "XADDPHR\0", "XADDPIX\0", "XADD0\0", "XADDINC\0" };
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,
- 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 /*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,
d1 = (a1f >> 3) & 0x07, d2 = (a2f >> 3) & 0x07,
WriteLog("dst: %s%s%s ", (cmd & 0x0008 ? "DSTEN " : ""), (cmd & 0x0010 ? "DSTENZ " : ""), (cmd & 0x0020 ? "DSTWRZ" : ""));
WriteLog("misc: %s%s ", (cmd & 0x0040 ? "CLIP_A1 " : ""), (cmd & 0x0080 ? "???" : ""));
WriteLog("a1ctl: %s%s%s ", (cmd & 0x0100 ? "UPDA1F " : ""), (cmd & 0x0200 ? "UPDA1 " : ""), (cmd & 0x0400 ? "UPDA2" : ""));
- WriteLog("mode: %s%s%s ", (cmd & 0x0800 ? "DSTA2 " : ""), (cmd & 0x1000 ? "GOURD " : ""), (cmd & 0x2000 ? "ZBUFF" : ""));
- WriteLog("ity: %s%s%s ", (cmd & 0x4000 ? "TOPBEN " : ""), (cmd & 0x8000 ? "TOPNEN " : ""), (cmd & 0x00010000 ? "PATDSEL" : ""));
+ WriteLog("mode: %s%s%s ", (cmd & 0x0800 ? "DSTA2 " : ""), (cmd & 0x1000 ? "GOURD " : ""), (cmd & 0x2000 ? "GOURZ" : ""));
+ WriteLog("ity: %s%s%s%s ", (cmd & 0x4000 ? "TOPBEN " : ""), (cmd & 0x8000 ? "TOPNEN " : ""), (cmd & 0x00010000 ? "PATDSEL" : ""), (cmd & 0x00020000 ? "ADDDSEL" : ""));
WriteLog("z-op: %s%s%s ", (cmd & 0x00040000 ? "ZMODELT " : ""), (cmd & 0x00080000 ? "ZMODEEQ " : ""), (cmd & 0x00100000 ? "ZMODEGT" : ""));
WriteLog("op: %s ", opStr[(cmd >> 21) & 0x0F]);
WriteLog("ctrl: %s%s%s%s%s%s\n", (cmd & 0x02000000 ? "CMPDST " : ""), (cmd & 0x04000000 ? "BCOMPEN " : ""), (cmd & 0x08000000 ? "DCOMPEN " : ""), (cmd & 0x10000000 ? "BKGWREN " : ""), (cmd & 0x20000000 ? "BUSHI " : ""), (cmd & 0x40000000 ? "SRCSHADE" : ""));
+ if (UPDA1)
+ WriteLog(" A1 step values: %d (X), %d (Y)\n", a1_step_x >> 16, a1_step_y >> 16);
+
if (UPDA2)
- {
WriteLog(" A2 step values: %d (X), %d (Y) [mask (%sused): %08X - %08X/%08X]\n", a2_step_x >> 16, a2_step_y >> 16, (a2f & 0x8000 ? "" : "un"), REG(A2_MASK), a2_mask_x, a2_mask_y);
- }
WriteLog(" A1 -> pitch: %d phrases, depth: %s, z-off: %d, width: %d (%02X), addctl: %s %s %s %s\n", 1 << p1, bppStr[d1], zo1, iw1, w1, ctrlStr[ac1&0x03], (ac1&0x04 ? "YADD1" : "YADD0"), (ac1&0x08 ? "XSIGNSUB" : "XSIGNADD"), (ac1&0x10 ? "YSIGNSUB" : "YSIGNADD"));
WriteLog(" A2 -> pitch: %d phrases, depth: %s, z-off: %d, width: %d (%02X), addctl: %s %s %s %s\n", 1 << p2, bppStr[d2], zo2, iw2, w2, ctrlStr[ac2&0x03], (ac2&0x04 ? "YADD1" : "YADD0"), (ac2&0x08 ? "XSIGNSUB" : "XSIGNADD"), (ac2&0x10 ? "YSIGNSUB" : "YSIGNADD"));
- WriteLog(" x/y: %d/%d\n", a2_x >> 16, a2_y >> 16);
+ WriteLog(" A1 x/y: %d/%d, A2 x/y: %d/%d Pattern: %08X%08X SRCDATA: %08X%08X\n", a1_x >> 16, a1_y >> 16, a2_x >> 16, a2_y >> 16, REG(PATTERNDATA), REG(PATTERNDATA + 4), REG(SRCDATA), REG(SRCDATA + 4));
// blit_start_log = 0;
// op_start_log = 1;
}
blitter_working = 1;
-#ifndef USE_GENERIC_BLITTER
- if (!blitter_execute_cached_code(blitter_in_cache(cmd)))
-#endif
- blitter_generic(cmd);
+//#ifndef USE_GENERIC_BLITTER
+// if (!blitter_execute_cached_code(blitter_in_cache(cmd)))
+//#endif
+ blitter_generic(cmd);
/*if (blit_start_log)
{
{
WriteLog("\nBytes at 004D58:\n");
for(int i=0x004D58; i<0x004D58+(10*127*4); i++)
- WriteLog("%02X ", jaguar_byte_read(i));
+ WriteLog("%02X ", JaguarReadByte(i));
WriteLog("\nBytes at F03000:\n");
for(int i=0xF03000; i<0xF03000+(6*127*4); i++)
- WriteLog("%02X ", jaguar_byte_read(i));
+ WriteLog("%02X ", JaguarReadByte(i));
WriteLog("\n\n");
}
}//*/
blitter_working = 0;
}
-uint32 blitter_long_read(uint32 offset)
+void blitter_init(void)
{
- return (blitter_word_read(offset) << 16) | blitter_word_read(offset+2);
+ blitter_reset();
}
-void blitter_long_write(uint32 offset, uint32 data)
+void blitter_reset(void)
{
- blitter_word_write(offset, data >> 16);
- blitter_word_write(offset+2, data & 0xFFFF);
+ memset(blitter_ram, 0x00, 0xA0);
}
-void blitter_init(void)
+void blitter_done(void)
{
- if (!blitter_cache_init)
- {
- for (int i=0;i<256;i++)
- {
- blitter_cache[i] = (struct s_blitter_cache *)malloc(sizeof(struct s_blitter_cache));
- blitter_cache[i]->next=null;
- blitter_cache[i]->prev=null;
- }
- blitter_cache_init = 1;
- }
-#ifndef USE_GENERIC_BLITTER
- #include "include/blit_i.h"
-#endif
+ WriteLog("BLIT: Done.\n");
+}
- blitter_reset();
-#ifdef GEN_CODE
- blitters_code_fp = fopen("include/blit_c.h", "awrt");
- blitters_code_init_fp = fopen("include/blit_i.h", "awrt");
-#endif
+uint8 BlitterReadByte(uint32 offset, uint32 who/*=UNKNOWN*/)
+{
+ offset &= 0xFF;
+
+ // status register
+//This isn't cycle accurate--how to fix? !!! FIX !!!
+ if (offset == (0x38 + 3))
+ return 0x01; // always idle
+
+//Attempted fix for AvP:
+ if (offset >= 0x04 && offset <= 0x07)
+// return (offset > 0x05 ? blitter_ram[PIXLINECOUNTER + offset - 0x04] : 0x00);
+// return 0x00; // WO register! What does it expect to see here???
+//This is it. I wonder if it just ignores the lower three bits?
+ return blitter_ram[A1_PIXEL + offset - 0x04];
+
+ return blitter_ram[offset];
}
-void blitter_reset(void)
+//Crappy!
+uint16 BlitterReadWord(uint32 offset, uint32 who/*=UNKNOWN*/)
{
- memset(blitter_ram, 0x00, 0xA0);
+ return ((uint16)BlitterReadByte(offset, who) << 8) | (uint16)BlitterReadByte(offset+1, who);
}
-void blitter_done(void)
+//Crappy!
+uint32 BlitterReadLong(uint32 offset, uint32 who/*=UNKNOWN*/)
{
-// blitter_list();
-#ifdef GEN_CODE
- fclose(blitters_code_fp);
- fclose(blitters_code_init_fp);
-#endif
- WriteLog("BLIT: Done.\n");
+ return (BlitterReadWord(offset, who) << 16) | BlitterReadWord(offset+2, who);
}
-void blitter_byte_write(uint32 offset, uint8 data)
+void BlitterWriteByte(uint32 offset, uint8 data, uint32 who/*=UNKNOWN*/)
{
/*if (offset & 0xFF == 0x7B)
-{
- WriteLog("--> Wrote to B_STOP: value -> %02X\n", data);
-}*/
+ WriteLog("--> Wrote to B_STOP: value -> %02X\n", data);*/
offset &= 0xFF;
// if ((offset >= 0x7C) && (offset <= 0x9B))
- if ((offset >= 0x7C) && (offset <= 0x8B))
+ // This handles writes to INTENSITY0-3 by also writing them to their proper places in
+ // PATTERNDATA & SOURCEDATA (should do the same for the Z registers! !!! FIX !!! [DONE])
+ if ((offset >= 0x7C) && (offset <= 0x9B))//8B))
{
switch (offset)
{
+ // INTENSITY registers 0-3
case 0x7C: break;
- case 0x7D: blitter_ram[0x69] = data; break;
- case 0x7E: blitter_ram[0x40] = data; break;
- case 0x7F: blitter_ram[0x41] = data; break;
+ case 0x7D: blitter_ram[PATTERNDATA + 1] = data; break;
+ case 0x7E: blitter_ram[SRCDATA + 0] = data; break;
+ case 0x7F: blitter_ram[SRCDATA + 1] = data; break;
case 0x80: break;
- case 0x81: blitter_ram[0x6B] = data; break;
- case 0x82: blitter_ram[0x42] = data; break;
- case 0x83: blitter_ram[0x43] = data; break;
+ case 0x81: blitter_ram[PATTERNDATA + 3] = data; break;
+ case 0x82: blitter_ram[SRCDATA + 2] = data; break;
+ case 0x83: blitter_ram[SRCDATA + 3] = data; break;
case 0x84: break;
- case 0x85: blitter_ram[0x6D] = data; break;
- case 0x86: blitter_ram[0x44] = data; break;
- case 0x87: blitter_ram[0x45] = data; break;
+ case 0x85: blitter_ram[PATTERNDATA + 5] = data; break;
+ case 0x86: blitter_ram[SRCDATA + 4] = data; break;
+ case 0x87: blitter_ram[SRCDATA + 5] = data; break;
case 0x88: break;
- case 0x89: blitter_ram[0x6F] = data; break;
-//Mistyped?
-// case 0x9A: blitter_ram[0x46] = data; break;
-// case 0x9B: blitter_ram[0x47] = data; break;
- case 0x8A: blitter_ram[0x46] = data; break;
- case 0x8B: blitter_ram[0x47] = data; break;
+ case 0x89: blitter_ram[PATTERNDATA + 7] = data; break;
+ case 0x8A: blitter_ram[SRCDATA + 6] = data; break;
+ case 0x8B: blitter_ram[SRCDATA + 7] = data; break;
+
+ // Z registers 0-3
+ case 0x8C: blitter_ram[SRCZINT + 0] = data; break;
+ case 0x8D: blitter_ram[SRCZINT + 1] = data; break;
+ case 0x8E: blitter_ram[SRCZFRAC + 0] = data; break;
+ case 0x8F: blitter_ram[SRCZFRAC + 1] = data; break;
+
+ case 0x90: blitter_ram[SRCZINT + 2] = data; break;
+ case 0x91: blitter_ram[SRCZINT + 3] = data; break;
+ case 0x92: blitter_ram[SRCZFRAC + 2] = data; break;
+ case 0x93: blitter_ram[SRCZFRAC + 3] = data; break;
+
+ case 0x94: blitter_ram[SRCZINT + 4] = data; break;
+ case 0x95: blitter_ram[SRCZINT + 5] = data; break;
+ case 0x96: blitter_ram[SRCZFRAC + 4] = data; break;
+ case 0x97: blitter_ram[SRCZFRAC + 5] = data; break;
+
+ case 0x98: blitter_ram[SRCZINT + 6] = data; break;
+ case 0x99: blitter_ram[SRCZINT + 7] = data; break;
+ case 0x9A: blitter_ram[SRCZFRAC + 6] = data; break;
+ case 0x9B: blitter_ram[SRCZFRAC + 7] = data; break;
}
}
blitter_ram[offset] = data;
}
-void blitter_word_write(uint32 offset, uint16 data)
+void BlitterWriteWord(uint32 offset, uint16 data, uint32 who/*=UNKNOWN*/)
{
- blitter_byte_write(offset+0, (data>>8) & 0xFF);
- blitter_byte_write(offset+1, data & 0xFF);
+//#if 1
+/* if (offset & 0xFF == A1_PIXEL && data == 14368)
+ {
+ WriteLog("\n1\nA1_PIXEL written by %s (%u)...\n\n\n", whoName[who], data);
+extern bool doGPUDis;
+doGPUDis = true;
+ }
+ if ((offset & 0xFF) == (A1_PIXEL + 2) && data == 14368)
+ {
+ WriteLog("\n2\nA1_PIXEL written by %s (%u)...\n\n\n", whoName[who], data);
+extern bool doGPUDis;
+doGPUDis = true;
+ }//*/
+//#endif
+
+ BlitterWriteByte(offset+0, (data>>8) & 0xFF, who);
+ BlitterWriteByte(offset+1, data & 0xFF, who);
if ((offset & 0xFF) == 0x3A)
// I.e., the second write of 32-bit value--not convinced this is the best way to do this!
// But then again, according to the Jaguar docs, this is correct...!
+{
+/*extern int blit_start_log;
+extern bool doGPUDis;
+if (blit_start_log)
+{
+ WriteLog("BLIT: Blitter started by %s...\n", whoName[who]);
+ doGPUDis = true;
+}//*/
blitter_blit(GET32(blitter_ram, 0x38));
-// Testing purposes only!
-//This does the clipping correctly, but not the Gouraud shading...
-// blitter2_exec(GET32(blitter_ram, 0x38));
+}
}
//F02278,9,A,B
-uint8 blitter_byte_read(uint32 offset)
+void BlitterWriteLong(uint32 offset, uint32 data, uint32 who/*=UNKNOWN*/)
{
- offset &= 0xFF;
-
- // status register
- if (offset == (0x38 + 3))
- return 0x01; // always idle
+//#if 1
+/* if ((offset & 0xFF) == A1_PIXEL && (data & 0xFFFF) == 14368)
+ {
+ WriteLog("\n3\nA1_PIXEL written by %s (%u)...\n\n\n", whoName[who], data);
+extern bool doGPUDis;
+doGPUDis = true;
+ }//*/
+//#endif
- return blitter_ram[offset];
+ BlitterWriteWord(offset, data >> 16, who);
+ BlitterWriteWord(offset+2, data & 0xFFFF, who);
}
-uint16 blitter_word_read(uint32 offset)
+void LogBlit(void)
{
- return ((uint16)blitter_byte_read(offset) << 8) | (uint16)blitter_byte_read(offset+1);
+ uint32 cmd = GET32(blitter_ram, 0x38);
+
+ WriteLog("Blit!\n");
+ WriteLog(" cmd = %08X\n", cmd);
+ WriteLog(" a1_base = %08X\n", a1_addr);
+ WriteLog(" a1_pitch = %d\n", a1_pitch);
+ WriteLog(" a1_psize = %d\n", a1_psize);
+ WriteLog(" a1_width = %d\n", a1_width);
+ WriteLog(" a1_xadd = %f (phrase=%d)\n", (float)a1_xadd / 65536.0, a1_phrase_mode);
+ WriteLog(" a1_yadd = %f\n", (float)a1_yadd / 65536.0);
+ WriteLog(" a1_xstep = %f\n", (float)a1_step_x / 65536.0);
+ WriteLog(" a1_ystep = %f\n", (float)a1_step_y / 65536.0);
+ WriteLog(" a1_x = %f\n", (float)a1_x / 65536.0);
+ WriteLog(" a1_y = %f\n", (float)a1_y / 65536.0);
+ WriteLog(" a1_zoffs = %i\n",a1_zoffs);
+
+ WriteLog(" a2_base = %08X\n", a2_addr);
+ WriteLog(" a2_pitch = %d\n", a2_pitch);
+ WriteLog(" a2_psize = %d\n", a2_psize);
+ WriteLog(" a2_width = %d\n", a2_width);
+ WriteLog(" a2_xadd = %f (phrase=%d)\n", (float)a2_xadd / 65536.0, a2_phrase_mode);
+ WriteLog(" a2_yadd = %f\n", (float)a2_yadd / 65536.0);
+ WriteLog(" a2_xstep = %f\n", (float)a2_step_x / 65536.0);
+ WriteLog(" a2_ystep = %f\n", (float)a2_step_y / 65536.0);
+ WriteLog(" a2_x = %f\n", (float)a2_x / 65536.0);
+ WriteLog(" a2_y = %f\n", (float)a2_y / 65536.0);
+ WriteLog(" a2_mask_x= 0x%.4x\n",a2_mask_x);
+ WriteLog(" a2_mask_y= 0x%.4x\n",a2_mask_y);
+ WriteLog(" a2_zoffs = %i\n",a2_zoffs);
+
+ WriteLog(" count = %d x %d\n", n_pixels, n_lines);
+
+ WriteLog(" COMMAND = %08X\n", cmd);
+ WriteLog(" DSTEN = %s\n", (DSTEN ? "1" : "0"));
+ WriteLog(" SRCEN = %s\n", (SRCEN ? "1" : "0"));
+ WriteLog(" PATDSEL = %s\n", (PATDSEL ? "1" : "0"));
+ WriteLog(" COLOR = %08X\n", REG(PATTERNDATA));
+ WriteLog(" DCOMPEN = %s\n", (DCOMPEN ? "1" : "0"));
+ WriteLog(" BCOMPEN = %s\n", (BCOMPEN ? "1" : "0"));
+ WriteLog(" CMPDST = %s\n", (CMPDST ? "1" : "0"));
+ WriteLog(" GOURZ = %s\n", (GOURZ ? "1" : "0"));
+ WriteLog(" GOURD = %s\n", (GOURD ? "1" : "0"));
+ WriteLog(" SRCSHADE = %s\n", (SRCSHADE ? "1" : "0"));
}
projects / virtualjaguar / blobdiff