//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));
if (specialLog)
{
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);
}
+/* if (BCOMPEN)
+ {
+ if (DSTA2)
+ a1_xadd = 0;
+ else
+ a2_xadd = 0;
+ }//*/
+
while (outer_loop--)
{
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;
-//Kludge for Hover Strike...
-//I wonder if this kludge is in conjunction with the SRCENX down below...
-if (BCOMPEN && SRCENX)
-{
- if (n_pixels < 8)
- bitPos = 8 - n_pixels;
-}
+
+ //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)
+ {
+ if (n_pixels < bppSrc)
+ bitPos = bppSrc - n_pixels;
+ }
inner_loop = n_pixels;
while (inner_loop--)
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;
-//Kludge...
-//Doesn't work...
-srcdata = 0xFFFFFFFF;
if (!DSTA2) // Data movement: A1 <- A2
{
// load src data and Z
// if (SRCEN)
- if (SRCEN || SRCENX) // Not sure if this is correct...
+ if (SRCEN || SRCENX) // Not sure if this is correct... (seems to be...!)
{
srcdata = READ_PIXEL(a2, REG(A2_FLAGS));
if (SRCENZ)
else // Use SRCDATA register...
{
srcdata = READ_RDATA(SRCDATA, a2, REG(A2_FLAGS), a2_phrase_mode);
- if (cmd & 0x0001C020) // PATDSEL | TOPBEN | TOPNEN | DSTWRZ
+ if (cmd & 0x0001C020) // PATDSEL | TOPBEN | TOPNEN | DSTWRZ
srczdata = READ_RDATA(SRCZINT, a2, REG(A2_FLAGS), a2_phrase_mode);
}
//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)
-// srcdata = READ_RDATA_1(SRCDATA, a2, a2_phrase_mode);
-// srcdata = READ_PIXEL_1(a2);
-// srcdata = READ_PIXEL(a2, REG(A2_FLAGS));
-/*
-#define PIXEL_SHIFT_1(a) (((~a##_x) >> 16) & 0x07)
-#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) ((JaguarReadByte(a##_addr+PIXEL_OFFSET_1(a), BLITTER) >> PIXEL_SHIFT_1(a)) & 0x01)
-
-#define PIXEL_SHIFT_1(a)
- (((~a2_x) >> 16) & 0x07)
-#define PIXEL_OFFSET_1(a)
- (((((UINT32)a2_y >> 16) * a2_width / 8) + (((UINT32)a2_x >> 19) & ~7)) * (1 + a2_pitch) + (((UINT32)a2_x >> 19) & 7))
-#define READ_PIXEL_1(a)
- ((JaguarReadByte(a2_addr+PIXEL_OFFSET_1(a2), BLITTER) >> PIXEL_SHIFT_1(a2)) & 0x01)
-*/
//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 = ((~(a2_x * 8)) >> 16) & 0x07;
-// uint32 pixShift = (~(a2_x >> 13)) & 0x07;
- uint32 pixShift = (~bitPos) & 0x07; // 8BPP only...!
-// uint32 pixOffset = (((((UINT32)a2_y >> 16) * a2_width / 8) + (((UINT32)(a2_x * 8) >> 19) & ~0x07)) * (1 + a2_pitch) + (((UINT32)(a2_x * 8) >> 19) & 0x07));
-// uint32 pixOffset = (((((UINT32)a2_y >> 16) * a2_width) + (((UINT32)(a2_x * 8) >> 19) & ~0x07)) * (1 + a2_pitch) + (((UINT32)(a2_x * 8) >> 19) & 0x07));
-// uint32 pixOffset = (((((UINT32)a2_y >> 16) * a2_width) + (((UINT32)a2_x >> 16) & ~0x07)) * (1 + a2_pitch)) + (((UINT32)a2_x >> 16) & 0x07);
-// srcdata = (JaguarReadByte(a2_addr+pixOffset, BLITTER) >> pixShift) & 0x01;
+ uint32 pixShift = (~bitPos) & (bppSrc - 1);
srcdata = (srcdata >> pixShift) & 0x01;
bitPos++;
+// if (bitPos % bppSrc == 0)
+// a2_x += 0x00010000;
}
/*
Interesting (Hover Strike--large letter):
WRITE_ZDATA(a1, REG(A1_FLAGS), srczdata);
}
}
- else // if (DSTA2)
+ else // if (DSTA2) // Data movement: A1 -> A2
{
// load src data and Z
if (SRCEN)
}
}
- // update x and y (inner loop)
-//kludge to test shiatsu...
-//nope, no wok.
-//Now it does! But crappy, crappy, crappy! !!! FIX !!!
-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_x += a1_xadd, a1_y += a1_yadd;
- a2_x = (a2_x + (a2_xadd / 8)) & 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;
}
}
-//more kludge to test shiatsu...
-//nope, no wok.
-//Now it does! But crappy, crappy, crappy! !!! FIX !!!
-if (BCOMPEN)
-{
- if (a2_x & 0xFFFF)
- a2_x = (a2_x & 0xFFFF0000) + 0x00010000;
-}//*/
-/*if (BCOMPEN && !DSTA2)//kludge for Hover Strike... !!! FIX !!!
-{
- a2_x -= n_pixels << 16;
- a2_x += ((n_pixels << 16) / 8) & 0XFFFF0000;
- a2_x += (n_pixels % 8 == 0 ? 0 : 1 << 16);
-}//*/
+/*
+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;//*/
//New: Phrase mode taken into account! :-p
- if (a1_phrase_mode)
+/* if (a1_phrase_mode) // v1
{
// Bump the pointer to the next phrase boundary
// Even though it works, this is crappy... Clean it up!
uint32 newxrem = (a1_x >> 16) % size;
a1_x &= 0x0000FFFF;
a1_x |= (((newx + (newxrem == 0 ? 0 : 1)) * size) & 0xFFFF) << 16;
+ }//*/
+ if (a1_phrase_mode) // v2
+ {
+ // 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)
+/* if (a2_phrase_mode) // v1
{
// Bump the pointer to the next phrase boundary
// Even though it works, this is crappy... Clean it up!
uint32 newx = (a2_x >> 16) / size;
uint32 newxrem = (a2_x >> 16) % size;
a2_x &= 0x0000FFFF;
- a2_x |= (((newx + (newxrem == 0 ? 0 : 1) /*+ extra*/) * size) & 0xFFFF) << 16;
+ 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_x += a1_step_x;
a1_y += a1_step_y;
a2_x += a2_step_x;
a2_y += a2_step_y;//*/
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;
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("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]);
BlitterWriteWord(offset, data >> 16, who);
BlitterWriteWord(offset+2, data & 0xFFFF, who);
}
+
+void LogBlit(void)
+{
+ 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 / commitdiff