PCem
changeset 103:354491040ce1
S3 ViRGE S3D emulation. Currently very slow, bugs, features missing.
Added S3 ViRGE/DX emulation, as the DX drivers are better than those for the original chip.
| author | TomW |
|---|---|
| date | Mon May 26 18:13:23 2014 +0100 |
| parents | 3d3d486d5fce |
| children | 41134e597e51 |
| files | src/ibm.h src/vid_s3_virge.c src/vid_s3_virge.h src/video.c src/win.c |
| diffstat | 5 files changed, 1208 insertions(+), 44 deletions(-) [+] |
line diff
1.1 --- a/src/ibm.h Sun May 11 14:09:13 2014 +0100 1.2 +++ b/src/ibm.h Mon May 26 18:13:23 2014 +0100 1.3 @@ -343,6 +343,7 @@ 1.4 #define GFX_OTI067 15 /*Oak OTI-067*/ 1.5 #define GFX_MACH64GX 16 /*ATI Graphics Pro Turbo (Mach64)*/ 1.6 #define GFX_CL_GD5429 17 /*Cirrus Logic CL-GD5429*/ 1.7 +#define GFX_VIRGEDX 18 /*S3 Virge/DX*/ 1.8 1.9 int gfxcard; 1.10
2.1 --- a/src/vid_s3_virge.c Sun May 11 14:09:13 2014 +0100 2.2 +++ b/src/vid_s3_virge.c Mon May 26 18:13:23 2014 +0100 2.3 @@ -1,6 +1,4 @@ 2.4 -/*S3 ViRGE emulation 2.5 - 2.6 - The SVGA core is largely the same as older S3 chips, but the blitter is totally different*/ 2.7 +/*S3 ViRGE emulation*/ 2.8 #include <stdlib.h> 2.9 #include "ibm.h" 2.10 #include "device.h" 2.11 @@ -12,7 +10,6 @@ 2.12 #include "vid_s3_virge.h" 2.13 #include "vid_svga.h" 2.14 #include "vid_svga_render.h" 2.15 -//#include "vid_sdac_ramdac.h" 2.16 2.17 typedef struct virge_t 2.18 { 2.19 @@ -34,7 +31,11 @@ 2.20 uint32_t linear_base, linear_size; 2.21 2.22 uint8_t pci_regs[256]; 2.23 - 2.24 + 2.25 + int is_375; 2.26 + 2.27 + int pixel_count, tri_count; 2.28 + 2.29 struct 2.30 { 2.31 uint32_t src_base; 2.32 @@ -69,6 +70,39 @@ 2.33 uint32_t pattern_8[8*8]; 2.34 uint32_t pattern_16[8*8]; 2.35 uint32_t pattern_32[8*8]; 2.36 + 2.37 + 2.38 + uint32_t z_base; 2.39 + uint32_t z_str; 2.40 + 2.41 + uint32_t tex_base; 2.42 + uint32_t tex_bdr_clr; 2.43 + uint32_t tbv, tbu; 2.44 + int32_t TdVdX, TdUdX; 2.45 + int32_t TdVdY, TdUdY; 2.46 + uint32_t tus, tvs; 2.47 + 2.48 + int32_t TdZdX, TdZdY; 2.49 + uint32_t tzs; 2.50 + 2.51 + int32_t TdWdX, TdWdY; 2.52 + uint32_t tws; 2.53 + 2.54 + int32_t TdDdX, TdDdY; 2.55 + uint32_t tds; 2.56 + 2.57 + int16_t TdGdX, TdBdX, TdRdX, TdAdX; 2.58 + int16_t TdGdY, TdBdY, TdRdY, TdAdY; 2.59 + uint32_t tgs, tbs, trs, tas; 2.60 + 2.61 + uint32_t TdXdY12; 2.62 + uint32_t txend12; 2.63 + uint32_t TdXdY01; 2.64 + uint32_t txend01; 2.65 + uint32_t TdXdY02; 2.66 + uint32_t txs; 2.67 + uint32_t tys; 2.68 + int ty01, ty12, tlr; 2.69 } s3d; 2.70 2.71 struct 2.72 @@ -104,6 +138,8 @@ 2.73 2.74 static void s3_virge_bitblt(virge_t *virge, int count, uint32_t cpu_dat); 2.75 2.76 +static void s3_virge_triangle(virge_t *virge); 2.77 + 2.78 static uint8_t s3_virge_mmio_read(uint32_t addr, void *p); 2.79 static uint16_t s3_virge_mmio_read_w(uint32_t addr, void *p); 2.80 static uint32_t s3_virge_mmio_read_l(uint32_t addr, void *p); 2.81 @@ -130,6 +166,8 @@ 2.82 CMD_SET_ITA_BYTE = (0 << 10), 2.83 CMD_SET_ITA_WORD = (1 << 10), 2.84 CMD_SET_ITA_DWORD = (2 << 10), 2.85 + 2.86 + CMD_SET_ZB_MODE = (3 << 24), 2.87 2.88 CMD_SET_XP = (1 << 25), 2.89 CMD_SET_YP = (1 << 26), 2.90 @@ -137,6 +175,10 @@ 2.91 CMD_SET_COMMAND_MASK = (15 << 27) 2.92 }; 2.93 2.94 +#define CMD_SET_ABC_SRC (1 << 18) 2.95 +#define CMD_SET_ABC_ENABLE (1 << 19) 2.96 +#define CMD_SET_TWE (1 << 26) 2.97 + 2.98 enum 2.99 { 2.100 CMD_SET_COMMAND_BITBLT = (0 << 27), 2.101 @@ -281,11 +323,12 @@ 2.102 { 2.103 virge_t *virge = (virge_t *)p; 2.104 svga_t *svga = &virge->svga; 2.105 + uint8_t ret; 2.106 2.107 if (((addr & 0xfff0) == 0x3d0 || (addr & 0xfff0) == 0x3b0) && !(svga->miscout & 1)) 2.108 addr ^= 0x60; 2.109 2.110 -// if (addr != 0x3da) pclog("S3 in %04X %04X:%08X\n", addr, CS, pc); 2.111 +// if (addr != 0x3da) pclog("S3 in %04X %04X:%08X ", addr, CS, pc); 2.112 switch (addr) 2.113 { 2.114 //case 0x3C6: case 0x3C7: case 0x3C8: case 0x3C9: 2.115 @@ -294,29 +337,38 @@ 2.116 2.117 case 0x3c5: 2.118 if (svga->seqaddr >= 0x10) 2.119 - return svga->seqregs[svga->seqaddr & 0x1f]; 2.120 + ret = svga->seqregs[svga->seqaddr & 0x1f]; 2.121 + else 2.122 + ret = svga_in(addr, svga); 2.123 break; 2.124 2.125 case 0x3D4: 2.126 - return svga->crtcreg; 2.127 + ret = svga->crtcreg; 2.128 + break; 2.129 case 0x3D5: 2.130 // pclog("Read CRTC R%02X %04X:%04X (%02x)\n", svga->crtcreg, CS, pc, svga->crtc[svga->crtcreg]); 2.131 switch (svga->crtcreg) 2.132 { 2.133 - case 0x2d: return virge->virge_id_high; /*Extended chip ID*/ 2.134 - case 0x2e: return virge->virge_id_low; /*New chip ID*/ 2.135 - case 0x2f: return virge->virge_rev; 2.136 - case 0x30: return virge->virge_id; /*Chip ID*/ 2.137 - case 0x31: return (svga->crtc[0x31] & 0xcf) | ((virge->ma_ext & 3) << 4); 2.138 - case 0x35: return (svga->crtc[0x35] & 0xf0) | (virge->bank & 0xf); 2.139 - case 0x36: return (svga->crtc[0x36] & 0xfc) | 2; /*PCI bus*/ 2.140 - case 0x51: return (svga->crtc[0x51] & 0xf0) | ((virge->bank >> 2) & 0xc) | ((virge->ma_ext >> 2) & 3); 2.141 - case 0x69: return virge->ma_ext; 2.142 - case 0x6a: return virge->bank; 2.143 + case 0x2d: ret = virge->virge_id_high; break; /*Extended chip ID*/ 2.144 + case 0x2e: ret = virge->virge_id_low; break; /*New chip ID*/ 2.145 + case 0x2f: ret = virge->virge_rev; break; 2.146 + case 0x30: ret = virge->virge_id; break; /*Chip ID*/ 2.147 + case 0x31: ret = (svga->crtc[0x31] & 0xcf) | ((virge->ma_ext & 3) << 4); break; 2.148 + case 0x35: ret = (svga->crtc[0x35] & 0xf0) | (virge->bank & 0xf); break; 2.149 + case 0x36: ret = (svga->crtc[0x36] & 0xfc) | 2; break; /*PCI bus*/ 2.150 + case 0x51: ret = (svga->crtc[0x51] & 0xf0) | ((virge->bank >> 2) & 0xc) | ((virge->ma_ext >> 2) & 3); break; 2.151 + case 0x69: ret = virge->ma_ext; break; 2.152 + case 0x6a: ret = virge->bank; break; 2.153 + default: ret = svga->crtc[svga->crtcreg]; break; 2.154 } 2.155 - return svga->crtc[svga->crtcreg]; 2.156 + break; 2.157 + 2.158 + default: 2.159 + ret = svga_in(addr, svga); 2.160 + break; 2.161 } 2.162 - return svga_in(addr, svga); 2.163 +// if (addr != 0x3da) pclog("%02X\n", ret); 2.164 + return ret; 2.165 } 2.166 2.167 static void s3_virge_recalctimings(svga_t *svga) 2.168 @@ -335,7 +387,7 @@ 2.169 if ((svga->crtc[0x67] & 0xc) != 0xc) /*VGA mode*/ 2.170 { 2.171 svga->ma_latch |= (virge->ma_ext << 16); 2.172 - 2.173 +pclog("VGA mode\n"); 2.174 if (svga->crtc[0x51] & 0x30) svga->rowoffset += (svga->crtc[0x51] & 0x30) << 4; 2.175 else if (svga->crtc[0x43] & 0x04) svga->rowoffset += 0x100; 2.176 if (!svga->rowoffset) svga->rowoffset = 256; 2.177 @@ -379,7 +431,9 @@ 2.178 svga->ma_latch = virge->streams.pri_fb1 >> 2; 2.179 else 2.180 svga->ma_latch = virge->streams.pri_fb0 >> 2; 2.181 - 2.182 + 2.183 + svga->hdisp = virge->streams.pri_w; 2.184 +pclog("Streams mode x_disp=%i\n", svga->hdisp); 2.185 svga->rowoffset = virge->streams.pri_stride >> 3; 2.186 2.187 switch ((virge->streams.pri_ctrl >> 24) & 0x7) 2.188 @@ -388,9 +442,11 @@ 2.189 svga->render = svga_render_8bpp_highres; 2.190 break; 2.191 case 3: /*KRGB-16 (1.5.5.5)*/ 2.192 + svga->htotal >>= 1; 2.193 svga->render = svga_render_15bpp_highres; 2.194 break; 2.195 case 5: /*RGB-16 (5.6.5)*/ 2.196 + svga->htotal >>= 1; 2.197 svga->render = svga_render_16bpp_highres; 2.198 break; 2.199 case 6: /*RGB-24 (8.8.8)*/ 2.200 @@ -536,7 +592,7 @@ 2.201 { 2.202 virge_t *virge = (virge_t *)p; 2.203 uint32_t ret = 0xffffffff; 2.204 -// pclog("New MMIO readl %08X\n", addr); 2.205 +// pclog("New MMIO readl %08X %04X(%08X):%08X ", addr, CS, cs, pc); 2.206 switch (addr & 0xfffc) 2.207 { 2.208 case 0x8180: 2.209 @@ -607,7 +663,7 @@ 2.210 break; 2.211 2.212 case 0x8504: 2.213 - ret = (0x1f << 8) | (1 << 13); 2.214 + ret = (0x10 << 8) | (1 << 13); 2.215 break; 2.216 case 0xa4d4: 2.217 ret = virge->s3d.src_base; 2.218 @@ -656,8 +712,9 @@ 2.219 break; 2.220 2.221 default: 2.222 - return s3_virge_mmio_read_w(addr, p) | (s3_virge_mmio_read_w(addr + 2, p) << 16); 2.223 + ret = s3_virge_mmio_read_w(addr, p) | (s3_virge_mmio_read_w(addr + 2, p) << 16); 2.224 } 2.225 +// pclog("%02x\n", ret); 2.226 return ret; 2.227 } 2.228 static void s3_virge_mmio_write(uint32_t addr, uint8_t val, void *p) 2.229 @@ -665,8 +722,8 @@ 2.230 virge_t *virge = (virge_t *)p; 2.231 svga_t *svga = &virge->svga; 2.232 2.233 -// pclog("New MMIO writeb %08X %02X\n", addr, val); 2.234 - 2.235 +// pclog("New MMIO writeb %08X %02X %04x(%08x):%08x\n", addr, val, CS, cs, pc); 2.236 + 2.237 if ((addr & 0xfffc) < 0x8000) 2.238 s3_virge_bitblt(virge, 8, val); 2.239 else switch (addr & 0xffff) 2.240 @@ -692,7 +749,7 @@ 2.241 static void s3_virge_mmio_write_w(uint32_t addr, uint16_t val, void *p) 2.242 { 2.243 virge_t *virge = (virge_t *)p; 2.244 -// pclog("New MMIO writew %08X %04X\n", addr, val); 2.245 +// pclog("New MMIO writew %08X %04X %04x(%08x):%08x\n", addr, val, CS, cs, pc); 2.246 if ((addr & 0xfffc) < 0x8000) 2.247 { 2.248 if (virge->s3d.cmd_set & CMD_SET_MS) 2.249 @@ -713,7 +770,7 @@ 2.250 virge_t *virge = (virge_t *)p; 2.251 svga_t *svga = &virge->svga; 2.252 // if ((addr & 0xfffc) >= 0x8000) 2.253 -// pclog("New MMIO writel %08X %08X\n", addr, val); 2.254 +// pclog("New MMIO writel %08X %08X %04x(%08x):%08x\n", addr, val, CS, cs, pc); 2.255 2.256 if ((addr & 0xfffc) < 0x8000) 2.257 { 2.258 @@ -760,6 +817,7 @@ 2.259 svga->fullchange = changeframecount; 2.260 break; 2.261 case 0x81cc: 2.262 + pclog("Write buffer_ctrl %08x\n", val); 2.263 virge->streams.buffer_ctrl = val; 2.264 s3_virge_recalctimings(svga); 2.265 break; 2.266 @@ -867,18 +925,18 @@ 2.267 case 0xa4d4: case 0xa8d4: 2.268 virge->s3d.src_base = val & 0x3ffff8; 2.269 break; 2.270 - case 0xa4d8: case 0xa8d8: 2.271 + case 0xa4d8: case 0xa8d8: case 0xb4d8: 2.272 virge->s3d.dest_base = val & 0x3ffff8; 2.273 break; 2.274 - case 0xa4dc: case 0xa8dc: 2.275 + case 0xa4dc: case 0xa8dc: case 0xb4dc: 2.276 virge->s3d.clip_l = (val >> 16) & 0x7ff; 2.277 virge->s3d.clip_r = val & 0x7ff; 2.278 break; 2.279 - case 0xa4e0: case 0xa8e0: 2.280 + case 0xa4e0: case 0xa8e0: case 0xb4e0: 2.281 virge->s3d.clip_t = (val >> 16) & 0x7ff; 2.282 virge->s3d.clip_b = val & 0x7ff; 2.283 break; 2.284 - case 0xa4e4: case 0xa8e4: 2.285 + case 0xa4e4: case 0xa8e4: case 0xb4e4: 2.286 virge->s3d.dest_str = (val >> 16) & 0xff8; 2.287 virge->s3d.src_str = val & 0xff8; 2.288 break; 2.289 @@ -938,6 +996,128 @@ 2.290 if (virge->s3d.cmd_set & CMD_SET_AE) 2.291 s3_virge_bitblt(virge, -1, 0); 2.292 break; 2.293 + 2.294 + case 0xb4d4: 2.295 + virge->s3d.z_base = val & 0x3ffff8; 2.296 + break; 2.297 + case 0xb4e8: 2.298 + virge->s3d.z_str = val & 0xff8; 2.299 + break; 2.300 + case 0xb4ec: 2.301 + virge->s3d.tex_base = val & 0x3ffff8; 2.302 + break; 2.303 + case 0xb4f0: 2.304 + virge->s3d.tex_bdr_clr = val & 0xffffff; 2.305 + break; 2.306 + case 0xb500: 2.307 + virge->s3d.cmd_set = val; 2.308 + if (!(val & CMD_SET_AE)) 2.309 + s3_virge_triangle(virge); 2.310 + break; 2.311 + case 0xb504: 2.312 + virge->s3d.tbv = val & 0xfffff; 2.313 + break; 2.314 + case 0xb508: 2.315 + virge->s3d.tbu = val & 0xfffff; 2.316 + break; 2.317 + case 0xb50c: 2.318 + virge->s3d.TdWdX = val; 2.319 + break; 2.320 + case 0xb510: 2.321 + virge->s3d.TdWdY = val; 2.322 + break; 2.323 + case 0xb514: 2.324 + virge->s3d.tws = val; 2.325 + break; 2.326 + case 0xb518: 2.327 + virge->s3d.TdDdX = val; 2.328 + break; 2.329 + case 0xb51c: 2.330 + virge->s3d.TdVdX = val; 2.331 + break; 2.332 + case 0xb520: 2.333 + virge->s3d.TdUdX = val; 2.334 + break; 2.335 + case 0xb524: 2.336 + virge->s3d.TdDdY = val; 2.337 + break; 2.338 + case 0xb528: 2.339 + virge->s3d.TdVdY = val; 2.340 + break; 2.341 + case 0xb52c: 2.342 + virge->s3d.TdUdY = val; 2.343 + break; 2.344 + case 0xb530: 2.345 + virge->s3d.tds = val; 2.346 + break; 2.347 + case 0xb534: 2.348 + virge->s3d.tvs = val; 2.349 + break; 2.350 + case 0xb538: 2.351 + virge->s3d.tus = val; 2.352 + break; 2.353 + case 0xb53c: 2.354 + virge->s3d.TdGdX = val >> 16; 2.355 + virge->s3d.TdBdX = val & 0xffff; 2.356 + break; 2.357 + case 0xb540: 2.358 + virge->s3d.TdAdX = val >> 16; 2.359 + virge->s3d.TdRdX = val & 0xffff; 2.360 + break; 2.361 + case 0xb544: 2.362 + virge->s3d.TdGdY = val >> 16; 2.363 + virge->s3d.TdBdY = val & 0xffff; 2.364 + break; 2.365 + case 0xb548: 2.366 + virge->s3d.TdAdY = val >> 16; 2.367 + virge->s3d.TdRdY = val & 0xffff; 2.368 + break; 2.369 + case 0xb54c: 2.370 + virge->s3d.tgs = (val >> 16) & 0xffff; 2.371 + virge->s3d.tbs = val & 0xffff; 2.372 + break; 2.373 + case 0xb550: 2.374 + virge->s3d.tas = (val >> 16) & 0xffff; 2.375 + virge->s3d.trs = val & 0xffff; 2.376 + break; 2.377 + 2.378 + case 0xb554: 2.379 + virge->s3d.TdZdX = val; 2.380 + break; 2.381 + case 0xb558: 2.382 + virge->s3d.TdZdY = val; 2.383 + break; 2.384 + case 0xb55c: 2.385 + virge->s3d.tzs = val; 2.386 + break; 2.387 + case 0xb560: 2.388 + virge->s3d.TdXdY12 = val; 2.389 + break; 2.390 + case 0xb564: 2.391 + virge->s3d.txend12 = val; 2.392 + break; 2.393 + case 0xb568: 2.394 + virge->s3d.TdXdY01 = val; 2.395 + break; 2.396 + case 0xb56c: 2.397 + virge->s3d.txend01 = val; 2.398 + break; 2.399 + case 0xb570: 2.400 + virge->s3d.TdXdY02 = val; 2.401 + break; 2.402 + case 0xb574: 2.403 + virge->s3d.txs = val; 2.404 + break; 2.405 + case 0xb578: 2.406 + virge->s3d.tys = val; 2.407 + break; 2.408 + case 0xb57c: 2.409 + virge->s3d.ty01 = (val >> 16) & 0x7ff; 2.410 + virge->s3d.ty12 = val & 0x7ff; 2.411 + virge->s3d.tlr = val >> 31; 2.412 + if (virge->s3d.cmd_set & CMD_SET_AE) 2.413 + s3_virge_triangle(virge); 2.414 + break; 2.415 } 2.416 } 2.417 2.418 @@ -958,6 +1138,10 @@ 2.419 } \ 2.420 } while (0) 2.421 2.422 +#define Z_READ(addr) *(uint16_t *)&vram[addr & 0x3fffff] 2.423 + 2.424 +#define Z_WRITE(addr, val) if (!(virge->s3d.cmd_set & CMD_SET_ZB_MODE)) *(uint16_t *)&vram[addr & 0x3fffff] = val 2.425 + 2.426 #define CLIP(x, y) \ 2.427 do \ 2.428 { \ 2.429 @@ -969,6 +1153,23 @@ 2.430 update = 0; \ 2.431 } while (0) 2.432 2.433 +#define Z_CLIP(Zzb, Zs) \ 2.434 + do \ 2.435 + { \ 2.436 + if (!(virge->s3d.cmd_set & CMD_SET_ZB_MODE)) \ 2.437 + switch ((virge->s3d.cmd_set >> 20) & 7) \ 2.438 + { \ 2.439 + case 0: update = 0; break; \ 2.440 + case 1: if (Zs <= Zzb) update = 0; else Zzb = Zs; break; \ 2.441 + case 2: if (Zs != Zzb) update = 0; else Zzb = Zs; break; \ 2.442 + case 3: if (Zs < Zzb) update = 0; else Zzb = Zs; break; \ 2.443 + case 4: if (Zs >= Zzb) update = 0; else Zzb = Zs; break; \ 2.444 + case 5: if (Zs == Zzb) update = 0; else Zzb = Zs; break; \ 2.445 + case 6: if (Zs > Zzb) update = 0; else Zzb = Zs; break; \ 2.446 + case 7: update = 1; Zzb = Zs; break; \ 2.447 + } \ 2.448 + } while (0) 2.449 + 2.450 #define MIX() \ 2.451 do \ 2.452 { \ 2.453 @@ -1089,11 +1290,16 @@ 2.454 virge->s3d.rop = (virge->s3d.cmd_set >> 17) & 0xff; 2.455 virge->s3d.data_left_count = 0; 2.456 2.457 - /* pclog("BitBlt start %i,%i %i,%i %02X\n", virge->s3d.dest_x, 2.458 +/* pclog("BitBlt start %i,%i %i,%i %i,%i %02X %x %x\n", 2.459 + virge->s3d.src_x, 2.460 + virge->s3d.src_y, 2.461 + virge->s3d.dest_x, 2.462 virge->s3d.dest_y, 2.463 virge->s3d.w, 2.464 virge->s3d.h, 2.465 - virge->s3d.rop);*/ 2.466 + virge->s3d.rop, 2.467 + virge->s3d.src_base, 2.468 + virge->s3d.dest_base);*/ 2.469 2.470 if (virge->s3d.cmd_set & CMD_SET_IDS) 2.471 return; 2.472 @@ -1214,11 +1420,11 @@ 2.473 virge->s3d.h = virge->s3d.r_height; 2.474 virge->s3d.rop = (virge->s3d.cmd_set >> 17) & 0xff; 2.475 2.476 -/* pclog("RctFll start %i,%i %i,%i %02X\n", virge->s3d.dest_x, 2.477 +/* pclog("RctFll start %i,%i %i,%i %02X %08x\n", virge->s3d.dest_x, 2.478 virge->s3d.dest_y, 2.479 virge->s3d.w, 2.480 virge->s3d.h, 2.481 - virge->s3d.rop);*/ 2.482 + virge->s3d.rop, virge->s3d.dest_base);*/ 2.483 } 2.484 2.485 while (count) 2.486 @@ -1315,6 +1521,859 @@ 2.487 } 2.488 } 2.489 2.490 +#define RGB15_TO_24(val, r, g, b) b = (val & 0x1f) << 3; \ 2.491 + g = (val & 0x3e0) >> 2; \ 2.492 + r = (val & 0x7c00) >> 7 2.493 + 2.494 +#define RGB24_TO_24(val, r, g, b) b = val & 0xff; \ 2.495 + g = (val & 0xff00) >> 8; \ 2.496 + r = (val & 0xff0000) >> 16 2.497 + 2.498 +#define RGB15(r, g, b) ((((b) >> 3) & 0x1f) | ((((g) >> 3) & 0x1f) << 5) | ((((r) >> 3) & 0x1f) << 10)) 2.499 + 2.500 +#define RGB24(r, g, b) ((b) | ((g) << 8) | ((r) << 16)) 2.501 + 2.502 +typedef struct s3d_state_t 2.503 +{ 2.504 + int32_t r, g, b, a, u, v, d, w; 2.505 + 2.506 + int32_t base_r, base_g, base_b, base_a, base_u, base_v, base_d, base_w; 2.507 + 2.508 + uint32_t base_z; 2.509 + 2.510 + uint32_t tbu, tbv; 2.511 + 2.512 + uint32_t cmd_set; 2.513 + int max_d; 2.514 + 2.515 + uint16_t *texture[10]; 2.516 + 2.517 + uint32_t tex_bdr_clr; 2.518 + 2.519 + int32_t x1, x2; 2.520 + int y; 2.521 +} s3d_state_t; 2.522 + 2.523 +typedef struct s3d_texture_state_t 2.524 +{ 2.525 + int level; 2.526 + int texture_shift; 2.527 + 2.528 + int32_t u, v; 2.529 +} s3d_texture_state_t; 2.530 + 2.531 +static void (*tex_read)(s3d_state_t *state, s3d_texture_state_t *texture_state, int *r_out, int *g_out, int *b_out, int *a_out); 2.532 +static void (*tex_sample)(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out); 2.533 +static void (*dest_pixel)(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out); 2.534 + 2.535 +#define MAX(a, b) ((a) > (b) ? (a) : (b)) 2.536 +#define MIN(a, b) ((a) < (b) ? (a) : (b)) 2.537 + 2.538 +static int _x, _y; 2.539 + 2.540 +static void tex_ARGB1555(s3d_state_t *state, s3d_texture_state_t *texture_state, int *r_out, int *g_out, int *b_out, int *a_out) 2.541 +{ 2.542 + int offset = ((texture_state->u & 0x7fc0000) >> texture_state->texture_shift) + 2.543 + (((texture_state->v & 0x7fc0000) >> texture_state->texture_shift) << texture_state->level); 2.544 + uint16_t val = state->texture[texture_state->level][offset]; 2.545 + 2.546 + if (((texture_state->u | texture_state->v) & 0xf8000000) == 0xf8000000 && !(state->cmd_set & CMD_SET_TWE)) 2.547 + val = state->tex_bdr_clr; 2.548 + 2.549 + *r_out = ((val & 0x7c00) >> 7) | ((val & 0x7000) >> 12); 2.550 + *g_out = ((val & 0x03e0) >> 2) | ((val & 0x0380) >> 7); 2.551 + *b_out = ((val & 0x001f) << 3) | ((val & 0x001c) >> 2); 2.552 + *a_out = (val & 0x8000) ? 0xff : 0; 2.553 +} 2.554 + 2.555 +static void tex_ARGB4444(s3d_state_t *state, s3d_texture_state_t *texture_state, int *r_out, int *g_out, int *b_out, int *a_out) 2.556 +{ 2.557 + int offset = ((texture_state->u & 0x7fc0000) >> texture_state->texture_shift) + 2.558 + (((texture_state->v & 0x7fc0000) >> texture_state->texture_shift) << texture_state->level); 2.559 + uint16_t val = state->texture[texture_state->level][offset]; 2.560 + 2.561 + if (((texture_state->u | texture_state->v) & 0xf8000000) == 0xf8000000 && !(state->cmd_set & CMD_SET_TWE)) 2.562 + val = state->tex_bdr_clr; 2.563 + 2.564 + *r_out = ((val & 0x0f00) >> 4) | ((val & 0x0f00) >> 8); 2.565 + *g_out = (val & 0x00f0) | ((val & 0x00f0) >> 4); 2.566 + *b_out = ((val & 0x000f) << 4) | (val & 0x000f); 2.567 + *a_out = ((val & 0xf000) >> 8) | ((val & 0xf000) >> 12); 2.568 + 2.569 +} 2.570 + 2.571 +static void tex_ARGB8888(s3d_state_t *state, s3d_texture_state_t *texture_state, int *r_out, int *g_out, int *b_out, int *a_out) 2.572 +{ 2.573 + int offset = ((texture_state->u & 0x7fc0000) >> texture_state->texture_shift) + 2.574 + (((texture_state->v & 0x7fc0000) >> texture_state->texture_shift) << texture_state->level); 2.575 + uint32_t val = ((uint32_t *)state->texture[texture_state->level])[offset]; 2.576 + 2.577 + if (((texture_state->u | texture_state->v) & 0xf8000000) == 0xf8000000 && !(state->cmd_set & CMD_SET_TWE)) 2.578 + val = state->tex_bdr_clr; 2.579 + 2.580 + *r_out = (val >> 16) & 0xff; 2.581 + *g_out = (val >> 8) & 0xff; 2.582 + *b_out = val & 0xff; 2.583 + *a_out = (val >> 24) & 0xff; 2.584 +} 2.585 + 2.586 +static void tex_sample_normal(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.587 +{ 2.588 + s3d_texture_state_t texture_state; 2.589 + 2.590 + texture_state.level = state->max_d; 2.591 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.592 + texture_state.u = state->u + state->tbu; 2.593 + texture_state.v = state->v + state->tbv; 2.594 + 2.595 + tex_read(state, &texture_state, r_out, g_out, b_out, a_out); 2.596 +} 2.597 + 2.598 +static void tex_sample_normal_filter(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.599 +{ 2.600 + s3d_texture_state_t texture_state; 2.601 + int tex_offset; 2.602 + int r[4], g[4], b[4], a[4]; 2.603 + int du, dv; 2.604 + int d[4]; 2.605 + 2.606 + texture_state.level = state->max_d; 2.607 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.608 + tex_offset = 1 << texture_state.texture_shift; 2.609 + 2.610 + texture_state.u = state->u + state->tbu; 2.611 + texture_state.v = state->v + state->tbv; 2.612 + tex_read(state, &texture_state, &r[0], &g[0], &b[0], &a[0]); 2.613 + du = (texture_state.u >> (texture_state.texture_shift - 8)) & 0xff; 2.614 + dv = (texture_state.v >> (texture_state.texture_shift - 8)) & 0xff; 2.615 + 2.616 + texture_state.u = state->u + state->tbu + tex_offset; 2.617 + texture_state.v = state->v + state->tbv; 2.618 + tex_read(state, &texture_state, &r[1], &g[1], &b[1], &a[1]); 2.619 + 2.620 + texture_state.u = state->u + state->tbu; 2.621 + texture_state.v = state->v + state->tbv + tex_offset; 2.622 + tex_read(state, &texture_state, &r[2], &g[2], &b[2], &a[2]); 2.623 + 2.624 + texture_state.u = state->u + state->tbu + tex_offset; 2.625 + texture_state.v = state->v + state->tbv + tex_offset; 2.626 + tex_read(state, &texture_state, &r[3], &g[3], &b[3], &a[3]); 2.627 + 2.628 + d[0] = (256 - du) * (256 - dv); 2.629 + d[1] = du * (256 - dv); 2.630 + d[2] = (256 - du) * dv; 2.631 + d[3] = du * dv; 2.632 + 2.633 + *r_out = (r[0] * d[0] + r[1] * d[1] + r[2] * d[2] + r[3] * d[3]) >> 16; 2.634 + *g_out = (g[0] * d[0] + g[1] * d[1] + g[2] * d[2] + g[3] * d[3]) >> 16; 2.635 + *b_out = (b[0] * d[0] + b[1] * d[1] + b[2] * d[2] + b[3] * d[3]) >> 16; 2.636 + *a_out = (a[0] * d[0] + a[1] * d[1] + a[2] * d[2] + a[3] * d[3]) >> 16; 2.637 +} 2.638 + 2.639 +static void tex_sample_mipmap(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.640 +{ 2.641 + s3d_texture_state_t texture_state; 2.642 + 2.643 + texture_state.level = MAX(MIN(9 - ((state->d >> 27) & 0xf), state->max_d), 0); 2.644 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.645 + texture_state.u = state->u + state->tbu; 2.646 + texture_state.v = state->v + state->tbv; 2.647 + 2.648 + tex_read(state, &texture_state, r_out, g_out, b_out, a_out); 2.649 +} 2.650 + 2.651 +static void tex_sample_mipmap_filter(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.652 +{ 2.653 + s3d_texture_state_t texture_state; 2.654 + int tex_offset; 2.655 + int r[4], g[4], b[4], a[4]; 2.656 + int du, dv; 2.657 + int d[4]; 2.658 + 2.659 + texture_state.level = MAX(MIN(9 - ((state->d >> 27) & 0xf), state->max_d), 0); 2.660 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.661 + tex_offset = 1 << texture_state.texture_shift; 2.662 + 2.663 + texture_state.u = state->u + state->tbu; 2.664 + texture_state.v = state->v + state->tbv; 2.665 + tex_read(state, &texture_state, &r[0], &g[0], &b[0], &a[0]); 2.666 + du = (texture_state.u >> (texture_state.texture_shift - 8)) & 0xff; 2.667 + dv = (texture_state.v >> (texture_state.texture_shift - 8)) & 0xff; 2.668 + 2.669 + texture_state.u = state->u + state->tbu + tex_offset; 2.670 + texture_state.v = state->v + state->tbv; 2.671 + tex_read(state, &texture_state, &r[1], &g[1], &b[1], &a[1]); 2.672 + 2.673 + texture_state.u = state->u + state->tbu; 2.674 + texture_state.v = state->v + state->tbv + tex_offset; 2.675 + tex_read(state, &texture_state, &r[2], &g[2], &b[2], &a[2]); 2.676 + 2.677 + texture_state.u = state->u + state->tbu + tex_offset; 2.678 + texture_state.v = state->v + state->tbv + tex_offset; 2.679 + tex_read(state, &texture_state, &r[3], &g[3], &b[3], &a[3]); 2.680 + 2.681 + d[0] = (256 - du) * (256 - dv); 2.682 + d[1] = du * (256 - dv); 2.683 + d[2] = (256 - du) * dv; 2.684 + d[3] = du * dv; 2.685 + 2.686 + *r_out = (r[0] * d[0] + r[1] * d[1] + r[2] * d[2] + r[3] * d[3]) >> 16; 2.687 + *g_out = (g[0] * d[0] + g[1] * d[1] + g[2] * d[2] + g[3] * d[3]) >> 16; 2.688 + *b_out = (b[0] * d[0] + b[1] * d[1] + b[2] * d[2] + b[3] * d[3]) >> 16; 2.689 + *a_out = (a[0] * d[0] + a[1] * d[1] + a[2] * d[2] + a[3] * d[3]) >> 16; 2.690 +} 2.691 + 2.692 +static void tex_sample_persp_normal(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.693 +{ 2.694 + s3d_texture_state_t texture_state; 2.695 + int32_t w = 0; 2.696 + 2.697 + if (state->w) 2.698 + w = (int32_t)(((1ULL << 27) << 19) / (int64_t)state->w); 2.699 + 2.700 + texture_state.level = state->max_d; 2.701 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.702 + texture_state.u = (int32_t)(((int64_t)state->u * (int64_t)w) >> (12 + state->max_d)) + state->tbu; 2.703 + texture_state.v = (int32_t)(((int64_t)state->v * (int64_t)w) >> (12 + state->max_d)) + state->tbv; 2.704 + 2.705 + tex_read(state, &texture_state, r_out, g_out, b_out, a_out); 2.706 +} 2.707 + 2.708 +static void tex_sample_persp_normal_filter(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.709 +{ 2.710 + s3d_texture_state_t texture_state; 2.711 + int32_t w = 0, u, v; 2.712 + int tex_offset; 2.713 + int r[4], g[4], b[4], a[4]; 2.714 + int du, dv; 2.715 + int d[4]; 2.716 + 2.717 + if (state->w) 2.718 + w = (int32_t)(((1ULL << 27) << 19) / (int64_t)state->w); 2.719 + 2.720 + u = (int32_t)(((int64_t)state->u * (int64_t)w) >> (12 + state->max_d)) + state->tbu; 2.721 + v = (int32_t)(((int64_t)state->v * (int64_t)w) >> (12 + state->max_d)) + state->tbv; 2.722 + 2.723 + texture_state.level = state->max_d; 2.724 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.725 + tex_offset = 1 << texture_state.texture_shift; 2.726 + 2.727 + texture_state.u = u; 2.728 + texture_state.v = v; 2.729 + tex_read(state, &texture_state, &r[0], &g[0], &b[0], &a[0]); 2.730 + du = (u >> (texture_state.texture_shift - 8)) & 0xff; 2.731 + dv = (v >> (texture_state.texture_shift - 8)) & 0xff; 2.732 + 2.733 + texture_state.u = u + tex_offset; 2.734 + texture_state.v = v; 2.735 + tex_read(state, &texture_state, &r[1], &g[1], &b[1], &a[1]); 2.736 + 2.737 + texture_state.u = u; 2.738 + texture_state.v = v + tex_offset; 2.739 + tex_read(state, &texture_state, &r[2], &g[2], &b[2], &a[2]); 2.740 + 2.741 + texture_state.u = u + tex_offset; 2.742 + texture_state.v = v + tex_offset; 2.743 + tex_read(state, &texture_state, &r[3], &g[3], &b[3], &a[3]); 2.744 + 2.745 + d[0] = (256 - du) * (256 - dv); 2.746 + d[1] = du * (256 - dv); 2.747 + d[2] = (256 - du) * dv; 2.748 + d[3] = du * dv; 2.749 + 2.750 + *r_out = (r[0] * d[0] + r[1] * d[1] + r[2] * d[2] + r[3] * d[3]) >> 16; 2.751 + *g_out = (g[0] * d[0] + g[1] * d[1] + g[2] * d[2] + g[3] * d[3]) >> 16; 2.752 + *b_out = (b[0] * d[0] + b[1] * d[1] + b[2] * d[2] + b[3] * d[3]) >> 16; 2.753 + *a_out = (a[0] * d[0] + a[1] * d[1] + a[2] * d[2] + a[3] * d[3]) >> 16; 2.754 +} 2.755 + 2.756 +static void tex_sample_persp_normal_375(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.757 +{ 2.758 + s3d_texture_state_t texture_state; 2.759 + int32_t w = 0; 2.760 + 2.761 + if (state->w) 2.762 + w = (int32_t)(((1ULL << 27) << 19) / (int64_t)state->w); 2.763 + 2.764 + texture_state.level = state->max_d; 2.765 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.766 + texture_state.u = (int32_t)(((int64_t)state->u * (int64_t)w) >> (8 + state->max_d)) + state->tbu; 2.767 + texture_state.v = (int32_t)(((int64_t)state->v * (int64_t)w) >> (8 + state->max_d)) + state->tbv; 2.768 + 2.769 + tex_read(state, &texture_state, r_out, g_out, b_out, a_out); 2.770 +} 2.771 + 2.772 +static void tex_sample_persp_normal_filter_375(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.773 +{ 2.774 + s3d_texture_state_t texture_state; 2.775 + int32_t w = 0, u, v; 2.776 + int tex_offset; 2.777 + int r[4], g[4], b[4], a[4]; 2.778 + int du, dv; 2.779 + int d[4]; 2.780 + 2.781 + if (state->w) 2.782 + w = (int32_t)(((1ULL << 27) << 19) / (int64_t)state->w); 2.783 + 2.784 + u = (int32_t)(((int64_t)state->u * (int64_t)w) >> (8 + state->max_d)) + state->tbu; 2.785 + v = (int32_t)(((int64_t)state->v * (int64_t)w) >> (8 + state->max_d)) + state->tbv; 2.786 + 2.787 + texture_state.level = state->max_d; 2.788 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.789 + tex_offset = 1 << texture_state.texture_shift; 2.790 + 2.791 + texture_state.u = u; 2.792 + texture_state.v = v; 2.793 + tex_read(state, &texture_state, &r[0], &g[0], &b[0], &a[0]); 2.794 + du = (u >> (texture_state.texture_shift - 8)) & 0xff; 2.795 + dv = (v >> (texture_state.texture_shift - 8)) & 0xff; 2.796 + 2.797 + texture_state.u = u + tex_offset; 2.798 + texture_state.v = v; 2.799 + tex_read(state, &texture_state, &r[1], &g[1], &b[1], &a[1]); 2.800 + 2.801 + texture_state.u = u; 2.802 + texture_state.v = v + tex_offset; 2.803 + tex_read(state, &texture_state, &r[2], &g[2], &b[2], &a[2]); 2.804 + 2.805 + texture_state.u = u + tex_offset; 2.806 + texture_state.v = v + tex_offset; 2.807 + tex_read(state, &texture_state, &r[3], &g[3], &b[3], &a[3]); 2.808 + 2.809 + d[0] = (256 - du) * (256 - dv); 2.810 + d[1] = du * (256 - dv); 2.811 + d[2] = (256 - du) * dv; 2.812 + d[3] = du * dv; 2.813 + 2.814 + *r_out = (r[0] * d[0] + r[1] * d[1] + r[2] * d[2] + r[3] * d[3]) >> 16; 2.815 + *g_out = (g[0] * d[0] + g[1] * d[1] + g[2] * d[2] + g[3] * d[3]) >> 16; 2.816 + *b_out = (b[0] * d[0] + b[1] * d[1] + b[2] * d[2] + b[3] * d[3]) >> 16; 2.817 + *a_out = (a[0] * d[0] + a[1] * d[1] + a[2] * d[2] + a[3] * d[3]) >> 16; 2.818 +} 2.819 + 2.820 + 2.821 +static void tex_sample_persp_mipmap(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.822 +{ 2.823 + s3d_texture_state_t texture_state; 2.824 + int32_t w = 0; 2.825 + 2.826 + if (state->w) 2.827 + w = (int32_t)(((1ULL << 27) << 19) / (int64_t)state->w); 2.828 + 2.829 + texture_state.level = MAX(MIN(9 - ((state->d >> 27) & 0xf), state->max_d), 0); 2.830 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.831 + texture_state.u = (int32_t)(((int64_t)state->u * (int64_t)w) >> (12 + state->max_d)) + state->tbu; 2.832 + texture_state.v = (int32_t)(((int64_t)state->v * (int64_t)w) >> (12 + state->max_d)) + state->tbv; 2.833 + 2.834 + tex_read(state, &texture_state, r_out, g_out, b_out, a_out); 2.835 +} 2.836 + 2.837 +static void tex_sample_persp_mipmap_filter(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.838 +{ 2.839 + s3d_texture_state_t texture_state; 2.840 + int32_t w = 0, u, v; 2.841 + int tex_offset; 2.842 + int r[4], g[4], b[4], a[4]; 2.843 + int du, dv; 2.844 + int d[4]; 2.845 + 2.846 + if (state->w) 2.847 + w = (int32_t)(((1ULL << 27) << 19) / (int64_t)state->w); 2.848 + 2.849 + u = (int32_t)(((int64_t)state->u * (int64_t)w) >> (12 + state->max_d)) + state->tbu; 2.850 + v = (int32_t)(((int64_t)state->v * (int64_t)w) >> (12 + state->max_d)) + state->tbv; 2.851 + 2.852 + texture_state.level = MAX(MIN(9 - ((state->d >> 27) & 0xf), state->max_d), 0); 2.853 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.854 + tex_offset = 1 << texture_state.texture_shift; 2.855 + 2.856 + texture_state.u = u; 2.857 + texture_state.v = v; 2.858 + tex_read(state, &texture_state, &r[0], &g[0], &b[0], &a[0]); 2.859 + du = (u >> (texture_state.texture_shift - 8)) & 0xff; 2.860 + dv = (v >> (texture_state.texture_shift - 8)) & 0xff; 2.861 + 2.862 + texture_state.u = u + tex_offset; 2.863 + texture_state.v = v; 2.864 + tex_read(state, &texture_state, &r[1], &g[1], &b[1], &a[1]); 2.865 + 2.866 + texture_state.u = u; 2.867 + texture_state.v = v + tex_offset; 2.868 + tex_read(state, &texture_state, &r[2], &g[2], &b[2], &a[2]); 2.869 + 2.870 + texture_state.u = u + tex_offset; 2.871 + texture_state.v = v + tex_offset; 2.872 + tex_read(state, &texture_state, &r[3], &g[3], &b[3], &a[3]); 2.873 + 2.874 + d[0] = (256 - du) * (256 - dv); 2.875 + d[1] = du * (256 - dv); 2.876 + d[2] = (256 - du) * dv; 2.877 + d[3] = du * dv; 2.878 + 2.879 + *r_out = (r[0] * d[0] + r[1] * d[1] + r[2] * d[2] + r[3] * d[3]) >> 16; 2.880 + *g_out = (g[0] * d[0] + g[1] * d[1] + g[2] * d[2] + g[3] * d[3]) >> 16; 2.881 + *b_out = (b[0] * d[0] + b[1] * d[1] + b[2] * d[2] + b[3] * d[3]) >> 16; 2.882 + *a_out = (a[0] * d[0] + a[1] * d[1] + a[2] * d[2] + a[3] * d[3]) >> 16; 2.883 +} 2.884 + 2.885 +static void tex_sample_persp_mipmap_375(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.886 +{ 2.887 + s3d_texture_state_t texture_state; 2.888 + int32_t w = 0; 2.889 + 2.890 + if (state->w) 2.891 + w = (int32_t)(((1ULL << 27) << 19) / (int64_t)state->w); 2.892 + 2.893 + texture_state.level = MAX(MIN(9 - ((state->d >> 27) & 0xf), state->max_d), 0); 2.894 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.895 + texture_state.u = (int32_t)(((int64_t)state->u * (int64_t)w) >> (8 + state->max_d)) + state->tbu; 2.896 + texture_state.v = (int32_t)(((int64_t)state->v * (int64_t)w) >> (8 + state->max_d)) + state->tbv; 2.897 + 2.898 + tex_read(state, &texture_state, r_out, g_out, b_out, a_out); 2.899 +} 2.900 + 2.901 +static void tex_sample_persp_mipmap_filter_375(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.902 +{ 2.903 + s3d_texture_state_t texture_state; 2.904 + int32_t w = 0, u, v; 2.905 + int tex_offset; 2.906 + int r[4], g[4], b[4], a[4]; 2.907 + int du, dv; 2.908 + int d[4]; 2.909 + 2.910 + if (state->w) 2.911 + w = (int32_t)(((1ULL << 27) << 19) / (int64_t)state->w); 2.912 + 2.913 + u = (int32_t)(((int64_t)state->u * (int64_t)w) >> (8 + state->max_d)) + state->tbu; 2.914 + v = (int32_t)(((int64_t)state->v * (int64_t)w) >> (8 + state->max_d)) + state->tbv; 2.915 + 2.916 + texture_state.level = MAX(MIN(9 - ((state->d >> 27) & 0xf), state->max_d), 0); 2.917 + texture_state.texture_shift = 18 + (9 - texture_state.level); 2.918 + tex_offset = 1 << texture_state.texture_shift; 2.919 + 2.920 + texture_state.u = u; 2.921 + texture_state.v = v; 2.922 + tex_read(state, &texture_state, &r[0], &g[0], &b[0], &a[0]); 2.923 + du = (u >> (texture_state.texture_shift - 8)) & 0xff; 2.924 + dv = (v >> (texture_state.texture_shift - 8)) & 0xff; 2.925 + 2.926 + texture_state.u = u + tex_offset; 2.927 + texture_state.v = v; 2.928 + tex_read(state, &texture_state, &r[1], &g[1], &b[1], &a[1]); 2.929 + 2.930 + texture_state.u = u; 2.931 + texture_state.v = v + tex_offset; 2.932 + tex_read(state, &texture_state, &r[2], &g[2], &b[2], &a[2]); 2.933 + 2.934 + texture_state.u = u + tex_offset; 2.935 + texture_state.v = v + tex_offset; 2.936 + tex_read(state, &texture_state, &r[3], &g[3], &b[3], &a[3]); 2.937 + 2.938 + d[0] = (256 - du) * (256 - dv); 2.939 + d[1] = du * (256 - dv); 2.940 + d[2] = (256 - du) * dv; 2.941 + d[3] = du * dv; 2.942 + 2.943 + *r_out = (r[0] * d[0] + r[1] * d[1] + r[2] * d[2] + r[3] * d[3]) >> 16; 2.944 + *g_out = (g[0] * d[0] + g[1] * d[1] + g[2] * d[2] + g[3] * d[3]) >> 16; 2.945 + *b_out = (b[0] * d[0] + b[1] * d[1] + b[2] * d[2] + b[3] * d[3]) >> 16; 2.946 + *a_out = (a[0] * d[0] + a[1] * d[1] + a[2] * d[2] + a[3] * d[3]) >> 16; 2.947 +} 2.948 + 2.949 + 2.950 +#define CLAMP_RGBA(r, g, b, a) do \ 2.951 + { \ 2.952 + if ((r) < 0) \ 2.953 + r = 0; \ 2.954 + if ((r) > 0xff) \ 2.955 + r = 0xff; \ 2.956 + if ((g) < 0) \ 2.957 + g = 0; \ 2.958 + if ((g) > 0xff) \ 2.959 + g = 0xff; \ 2.960 + if ((b) < 0) \ 2.961 + b = 0; \ 2.962 + if ((b) > 0xff) \ 2.963 + b = 0xff; \ 2.964 + if ((a) < 0) \ 2.965 + a = 0; \ 2.966 + if ((a) > 0xff) \ 2.967 + a = 0xff; \ 2.968 + } \ 2.969 + while (0) 2.970 + 2.971 +#define CLAMP_RGB(r, g, b) do \ 2.972 + { \ 2.973 + if ((r) < 0) \ 2.974 + r = 0; \ 2.975 + if ((r) > 0xff) \ 2.976 + r = 0xff; \ 2.977 + if ((g) < 0) \ 2.978 + g = 0; \ 2.979 + if ((g) > 0xff) \ 2.980 + g = 0xff; \ 2.981 + if ((b) < 0) \ 2.982 + b = 0; \ 2.983 + if ((b) > 0xff) \ 2.984 + b = 0xff; \ 2.985 + } \ 2.986 + while (0) 2.987 + 2.988 +static void dest_pixel_gouraud_shaded_triangle(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.989 +{ 2.990 + *r_out = state->r >> 7; 2.991 + *g_out = state->g >> 7; 2.992 + *b_out = state->b >> 7; 2.993 + *a_out = state->a >> 7; 2.994 + CLAMP_RGBA(*r_out, *g_out, *b_out, *a_out); 2.995 +} 2.996 + 2.997 +static void dest_pixel_unlit_texture_triangle(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.998 +{ 2.999 + tex_sample(state, r_out, g_out, b_out, a_out); 2.1000 + 2.1001 + if (state->cmd_set & CMD_SET_ABC_SRC) 2.1002 + *a_out = state->a >> 7; 2.1003 +} 2.1004 + 2.1005 +static void dest_pixel_lit_texture_decal(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.1006 +{ 2.1007 + tex_sample(state, r_out, g_out, b_out, a_out); 2.1008 + 2.1009 + if (state->cmd_set & CMD_SET_ABC_SRC) 2.1010 + *a_out = state->a >> 7; 2.1011 +} 2.1012 + 2.1013 +static void dest_pixel_lit_texture_reflection(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.1014 +{ 2.1015 + int tex_r, tex_g, tex_b, tex_a; 2.1016 + 2.1017 + tex_sample(state, &tex_r, &tex_g, &tex_b, &tex_a); 2.1018 + 2.1019 + *r_out = state->r >> 7; 2.1020 + *g_out = state->g >> 7; 2.1021 + *b_out = state->b >> 7; 2.1022 + *a_out = state->a >> 7; 2.1023 + CLAMP_RGBA(*r_out, *g_out, *b_out, *a_out); 2.1024 + 2.1025 + *(r_out) += tex_r; 2.1026 + *(g_out) += tex_g; 2.1027 + *(b_out) += tex_b; 2.1028 + 2.1029 + CLAMP_RGB(*r_out, *g_out, *b_out); 2.1030 + 2.1031 + if (!(state->cmd_set & CMD_SET_ABC_SRC)) 2.1032 + *a_out = tex_a; 2.1033 +} 2.1034 + 2.1035 +static void dest_pixel_lit_texture_modulate(s3d_state_t *state, int *r_out, int *g_out, int *b_out, int *a_out) 2.1036 +{ 2.1037 + int r = state->r >> 7, g = state->g >> 7, b = state->b >> 7, a = state->a >> 7; 2.1038 + 2.1039 + tex_sample(state, r_out, g_out, b_out, a_out); 2.1040 + 2.1041 + CLAMP_RGBA(r, g, b, a); 2.1042 + 2.1043 + *r_out = ((*r_out) * r) >> 8; 2.1044 + *g_out = ((*g_out) * g) >> 8; 2.1045 + *b_out = ((*b_out) * b) >> 8; 2.1046 + 2.1047 + if (state->cmd_set & CMD_SET_ABC_SRC) 2.1048 + *a_out = a; 2.1049 +} 2.1050 + 2.1051 +static void tri(virge_t *virge, s3d_state_t *state, int yc, int32_t dx1, int32_t dx2) 2.1052 +{ 2.1053 + uint8_t *vram = virge->svga.vram; 2.1054 + 2.1055 + int x_dir = virge->s3d.tlr ? 1 : -1; 2.1056 + 2.1057 + int use_z = !(virge->s3d.cmd_set & CMD_SET_ZB_MODE); 2.1058 + 2.1059 + int y_count = yc; 2.1060 + 2.1061 + int bpp = 1; 2.1062 + 2.1063 + uint32_t dest_offset = virge->s3d.dest_base + (state->y * virge->s3d.dest_str); 2.1064 + uint32_t z_offset = virge->s3d.z_base + (state->y * virge->s3d.z_str); 2.1065 + 2.1066 + for (; y_count > 0; y_count--) 2.1067 + { 2.1068 + int x = state->x1 >> 20; 2.1069 + int xe = state->x2 >> 20; 2.1070 + uint32_t z = state->base_z; 2.1071 + if (x != xe && (x_dir > 0 && x < xe) || (x_dir < 0 && x > xe)) 2.1072 + { 2.1073 + int dx = (x_dir > 0) ? 8 - ((state->x1 >> 16) & 0xf) : ((state->x1 >> 16) & 0xf) - 8; 2.1074 + state->r = state->base_r + ((virge->s3d.TdRdX * dx) >> 4); 2.1075 + state->g = state->base_g + ((virge->s3d.TdGdX * dx) >> 4); 2.1076 + state->b = state->base_b + ((virge->s3d.TdBdX * dx) >> 4); 2.1077 + state->a = state->base_a + ((virge->s3d.TdAdX * dx) >> 4); 2.1078 + state->u = state->base_u + ((virge->s3d.TdUdX * dx) >> 4); 2.1079 + state->v = state->base_v + ((virge->s3d.TdVdX * dx) >> 4); 2.1080 + state->w = state->base_w + ((virge->s3d.TdWdX * dx) >> 4); 2.1081 + state->d = state->base_d + ((virge->s3d.TdDdX * dx) >> 4); 2.1082 + z += ((virge->s3d.TdZdX * dx) >> 4); 2.1083 +// pclog("Draw Y=%i X=%i to XE=%i %i %08x %08x %08x %08x %08x %08x %08x %08x %i %08x\n", state->y, x, xe, dx, state->x1, state->x2, dx1, virge->s3d.TdWdX, state->u, state->v, virge->s3d.TdUdX, virge->s3d.TdUdY, dx, (virge->s3d.TdUdX * dx) >> 4); 2.1084 + 2.1085 + for (; x != ((xe + x_dir) & 0xfff); x = (x + x_dir) & 0xfff) 2.1086 + { 2.1087 + uint32_t dest_addr = dest_offset + (x << bpp); 2.1088 + uint32_t z_addr = z_offset + (x << bpp); 2.1089 + int update = 1; 2.1090 + int16_t src_z; 2.1091 + _x = x; _y = state->y; 2.1092 + 2.1093 + if (use_z) 2.1094 + { 2.1095 + src_z = Z_READ(z_addr); 2.1096 + Z_CLIP(src_z, z >> 16); 2.1097 + } 2.1098 + CLIP(x, state->y); 2.1099 + 2.1100 + if (update) 2.1101 + { 2.1102 + int dest_r, dest_g, dest_b, dest_a; 2.1103 + uint32_t dest_col; 2.1104 + 2.1105 + dest_pixel(state, &dest_r, &dest_g, &dest_b, &dest_a); 2.1106 + 2.1107 + if (virge->s3d.cmd_set & CMD_SET_ABC_ENABLE) 2.1108 + { 2.1109 + uint32_t src_col; 2.1110 + int src_r, src_g, src_b; 2.1111 + 2.1112 + switch (bpp) 2.1113 + { 2.1114 + case 0: /*8 bpp*/ 2.1115 + /*Not implemented yet*/ 2.1116 + break; 2.1117 + case 1: /*16 bpp*/ 2.1118 + src_col = *(uint16_t *)&vram[dest_addr & 0x3fffff]; 2.1119 + RGB15_TO_24(src_col, src_r, src_g, src_b); 2.1120 + break; 2.1121 + case 2: /*24 bpp*/ 2.1122 + src_col = (*(uint32_t *)&vram[dest_addr & 0x3fffff]) & 0xffffff; 2.1123 + RGB24_TO_24(src_col, src_r, src_g, src_b); 2.1124 + break; 2.1125 + } 2.1126 + 2.1127 + dest_r = ((dest_r * dest_a) + (src_r * (255 - dest_a))) / 255; 2.1128 + dest_g = ((dest_g * dest_a) + (src_g * (255 - dest_a))) / 255; 2.1129 + dest_b = ((dest_b * dest_a) + (src_b * (255 - dest_a))) / 255; 2.1130 + } 2.1131 + 2.1132 + switch (bpp) 2.1133 + { 2.1134 + case 0: /*8 bpp*/ 2.1135 + /*Not implemented yet*/ 2.1136 + break; 2.1137 + case 1: /*16 bpp*/ 2.1138 + dest_col = RGB15(dest_r, dest_g, dest_b); 2.1139 + *(uint16_t *)&vram[dest_addr & 0x3fffff] = dest_col; 2.1140 + virge->svga.changedvram[(dest_addr & 0x3fffff) >> 12] = changeframecount; 2.1141 + break; 2.1142 + case 2: /*24 bpp*/ 2.1143 + dest_col = RGB24(dest_r, dest_g, dest_b); 2.1144 + *(uint32_t *)&vram[dest_addr & 0x3fffff] = dest_col; 2.1145 + virge->svga.changedvram[(dest_addr & 0x3fffff) >> 12] = changeframecount; 2.1146 + break; 2.1147 + } 2.1148 + 2.1149 + if (use_z) 2.1150 + Z_WRITE(z_addr, src_z); 2.1151 + } 2.1152 + 2.1153 + z += virge->s3d.TdZdX; 2.1154 + state->u += virge->s3d.TdUdX; 2.1155 + state->v += virge->s3d.TdVdX; 2.1156 + state->r += virge->s3d.TdRdX; 2.1157 + state->g += virge->s3d.TdGdX; 2.1158 + state->b += virge->s3d.TdBdX; 2.1159 + state->a += virge->s3d.TdAdX; 2.1160 + state->d += virge->s3d.TdDdX; 2.1161 + state->w += virge->s3d.TdWdX; 2.1162 + virge->pixel_count++; 2.1163 + } 2.1164 + } 2.1165 + state->x1 += dx1; 2.1166 + state->x2 += dx2; 2.1167 + state->base_u += virge->s3d.TdUdY; 2.1168 + state->base_v += virge->s3d.TdVdY; 2.1169 + state->base_z += virge->s3d.TdZdY; 2.1170 + state->base_r += virge->s3d.TdRdY; 2.1171 + state->base_g += virge->s3d.TdGdY; 2.1172 + state->base_b += virge->s3d.TdBdY; 2.1173 + state->base_a += virge->s3d.TdAdY; 2.1174 + state->base_d += virge->s3d.TdDdY; 2.1175 + state->base_w += virge->s3d.TdWdY; 2.1176 + state->y--; 2.1177 + dest_offset -= virge->s3d.dest_str; 2.1178 + z_offset -= virge->s3d.z_str; 2.1179 + } 2.1180 +} 2.1181 + 2.1182 +static int tex_size[8] = 2.1183 +{ 2.1184 + 4*2, 2.1185 + 2*2, 2.1186 + 2*2, 2.1187 + 1*2, 2.1188 + 2/1, 2.1189 + 2/1, 2.1190 + 1*2, 2.1191 + 1*2 2.1192 +}; 2.1193 + 2.1194 +static void s3_virge_triangle(virge_t *virge) 2.1195 +{ 2.1196 + s3d_state_t state; 2.1197 + 2.1198 + uint32_t tex_base; 2.1199 + int c; 2.1200 + 2.1201 + state.tbu = virge->s3d.tbu << 11; 2.1202 + state.tbv = virge->s3d.tbv << 11; 2.1203 + 2.1204 + state.max_d = (virge->s3d.cmd_set >> 8) & 15; 2.1205 + 2.1206 + state.tex_bdr_clr = virge->s3d.tex_bdr_clr; 2.1207 + 2.1208 + state.cmd_set = virge->s3d.cmd_set; 2.1209 + 2.1210 + state.base_u = virge->s3d.tus; 2.1211 + state.base_v = virge->s3d.tvs; 2.1212 + state.base_z = virge->s3d.tzs; 2.1213 + state.base_r = (int32_t)virge->s3d.trs; 2.1214 + state.base_g = (int32_t)virge->s3d.tgs; 2.1215 + state.base_b = (int32_t)virge->s3d.tbs; 2.1216 + state.base_a = (int32_t)virge->s3d.tas; 2.1217 + state.base_d = virge->s3d.tds; 2.1218 + state.base_w = virge->s3d.tws; 2.1219 + 2.1220 + tex_base = virge->s3d.tex_base; 2.1221 + for (c = 9; c >= 0; c--) 2.1222 + { 2.1223 + state.texture[c] = (uint16_t *)&virge->svga.vram[tex_base]; 2.1224 + if (c <= state.max_d) 2.1225 + tex_base += ((1 << (c*2)) * tex_size[(virge->s3d.cmd_set >> 5) & 7]) / 2; 2.1226 + } 2.1227 + 2.1228 + switch ((virge->s3d.cmd_set >> 27) & 0xf) 2.1229 + { 2.1230 + case 0: 2.1231 + dest_pixel = dest_pixel_gouraud_shaded_triangle; 2.1232 +// pclog("dest_pixel_gouraud_shaded_triangle\n"); 2.1233 + break; 2.1234 + case 1: 2.1235 + case 5: 2.1236 + switch ((virge->s3d.cmd_set >> 15) & 0x3) 2.1237 + { 2.1238 + case 0: 2.1239 + dest_pixel = dest_pixel_lit_texture_reflection; 2.1240 +// pclog("dest_pixel_lit_texture_reflection\n"); 2.1241 + break; 2.1242 + case 1: 2.1243 + dest_pixel = dest_pixel_lit_texture_modulate; 2.1244 +// pclog("dest_pixel_lit_texture_modulate\n"); 2.1245 + break; 2.1246 + case 2: 2.1247 + dest_pixel = dest_pixel_lit_texture_decal; 2.1248 +// pclog("dest_pixel_lit_texture_decal\n"); 2.1249 + break; 2.1250 + default: 2.1251 + pclog("bad triangle type %x\n", (virge->s3d.cmd_set >> 27) & 0xf); 2.1252 + return; 2.1253 + } 2.1254 + break; 2.1255 + case 2: 2.1256 + case 6: 2.1257 + dest_pixel = dest_pixel_unlit_texture_triangle; 2.1258 +// pclog("dest_pixel_unlit_texture_triangle\n"); 2.1259 + break; 2.1260 + default: 2.1261 + pclog("bad triangle type %x\n", (virge->s3d.cmd_set >> 27) & 0xf); 2.1262 + return; 2.1263 + } 2.1264 + 2.1265 + switch (((virge->s3d.cmd_set >> 12) & 7) | ((virge->s3d.cmd_set & (1 << 29)) ? 8 : 0)) 2.1266 + { 2.1267 + case 0: case 1: 2.1268 + tex_sample = tex_sample_mipmap; 2.1269 +// pclog("use tex_sample_mipmap\n"); 2.1270 + break; 2.1271 + case 2: case 3: 2.1272 + tex_sample = tex_sample_mipmap_filter; 2.1273 +// pclog("use tex_sample_mipmap_filter\n"); 2.1274 + break; 2.1275 + case 4: case 5: 2.1276 + tex_sample = tex_sample_normal; 2.1277 +// pclog("use tex_sample_normal\n"); 2.1278 + break; 2.1279 + case 6: case 7: 2.1280 + tex_sample = tex_sample_normal_filter; 2.1281 +// pclog("use tex_sample_normal_filter\n"); 2.1282 + break; 2.1283 + case (0 | 8): case (1 | 8): 2.1284 + if (virge->is_375) 2.1285 + tex_sample = tex_sample_persp_mipmap_375; 2.1286 + else 2.1287 + tex_sample = tex_sample_persp_mipmap; 2.1288 +// pclog("use tex_sample_persp_mipmap\n"); 2.1289 + break; 2.1290 + case (2 | 8): case (3 | 8): 2.1291 + if (virge->is_375) 2.1292 + tex_sample = tex_sample_persp_mipmap_filter_375; 2.1293 + else 2.1294 + tex_sample = tex_sample_persp_mipmap_filter; 2.1295 +// pclog("use tex_sample_persp_mipmap_filter\n"); 2.1296 + break; 2.1297 + case (4 | 8): case (5 | 8): 2.1298 + if (virge->is_375) 2.1299 + tex_sample = tex_sample_persp_normal_375; 2.1300 + else 2.1301 + tex_sample = tex_sample_persp_normal; 2.1302 +// pclog("use tex_sample_persp_normal\n"); 2.1303 + break; 2.1304 + case (6 | 8): case (7 | 8): 2.1305 + if (virge->is_375) 2.1306 + tex_sample = tex_sample_persp_normal_filter_375; 2.1307 + else 2.1308 + tex_sample = tex_sample_persp_normal_filter; 2.1309 +// pclog("use tex_sample_persp_normal_filter\n"); 2.1310 + break; 2.1311 + } 2.1312 + 2.1313 + switch ((virge->s3d.cmd_set >> 5) & 7) 2.1314 + { 2.1315 + case 0: 2.1316 + tex_read = tex_ARGB8888; 2.1317 + break; 2.1318 + case 1: 2.1319 + tex_read = tex_ARGB4444; 2.1320 +// pclog("tex_ARGB4444\n"); 2.1321 + break; 2.1322 + case 2: 2.1323 + tex_read = tex_ARGB1555; 2.1324 +// pclog("tex_ARGB1555 %i\n", (virge->s3d.cmd_set >> 5) & 7); 2.1325 + break; 2.1326 + default: 2.1327 + pclog("bad texture type %i\n", (virge->s3d.cmd_set >> 5) & 7); 2.1328 + tex_read = tex_ARGB1555; 2.1329 + } 2.1330 + 2.1331 +// pclog("Triangle %i %i,%i to %i,%i %08x\n", y, x1 >> 20, y, virge->s3d.txend01 >> 20, y - (virge->s3d.ty01 + virge->s3d.ty12), state.cmd_set); 2.1332 + 2.1333 + state.y = virge->s3d.tys; 2.1334 + state.x1 = virge->s3d.txs; 2.1335 + state.x2 = virge->s3d.txend01; 2.1336 + tri(virge, &state, virge->s3d.ty01, virge->s3d.TdXdY02, virge->s3d.TdXdY01); 2.1337 + state.x2 = virge->s3d.txend12; 2.1338 + tri(virge, &state, virge->s3d.ty12, virge->s3d.TdXdY02, virge->s3d.TdXdY12); 2.1339 + 2.1340 + virge->tri_count++; 2.1341 +} 2.1342 + 2.1343 2.1344 static void s3_virge_hwcursor_draw(svga_t *svga, int displine) 2.1345 { 2.1346 @@ -1434,13 +2493,13 @@ 2.1347 if (virge->pci_regs[0x30] & 0x01) 2.1348 { 2.1349 uint32_t addr = (virge->pci_regs[0x32] << 16) | (virge->pci_regs[0x33] << 24); 2.1350 - pclog("Virge bios_rom enabled at %08x\n", addr); 2.1351 +// pclog("Virge bios_rom enabled at %08x\n", addr); 2.1352 mem_mapping_set_addr(&virge->bios_rom.mapping, addr, 0x8000); 2.1353 mem_mapping_enable(&virge->bios_rom.mapping); 2.1354 } 2.1355 else 2.1356 { 2.1357 - pclog("Virge bios_rom disabled\n"); 2.1358 +// pclog("Virge bios_rom disabled\n"); 2.1359 mem_mapping_disable(&virge->bios_rom.mapping); 2.1360 } 2.1361 return; 2.1362 @@ -1512,6 +2571,80 @@ 2.1363 virge->svga.crtc[0x37] = 1;// | (7 << 5); 2.1364 virge->svga.crtc[0x53] = 1 << 3; 2.1365 virge->svga.crtc[0x59] = 0x70; 2.1366 + 2.1367 + virge->is_375 = 0; 2.1368 + 2.1369 + pci_add(s3_virge_pci_read, s3_virge_pci_write, virge); 2.1370 + 2.1371 + return virge; 2.1372 +} 2.1373 + 2.1374 +static void *s3_virge_375_init() 2.1375 +{ 2.1376 + virge_t *virge = malloc(sizeof(virge_t)); 2.1377 + memset(virge, 0, sizeof(virge_t)); 2.1378 + 2.1379 + svga_init(&virge->svga, virge, 1 << 22, /*4mb*/ 2.1380 + s3_virge_recalctimings, 2.1381 + s3_virge_in, s3_virge_out, 2.1382 + s3_virge_hwcursor_draw); 2.1383 + 2.1384 + rom_init(&virge->bios_rom, "roms/86c375_1.bin", 0xc0000, 0x8000, 0x7fff, 0, MEM_MAPPING_EXTERNAL); 2.1385 + if (PCI) 2.1386 + mem_mapping_disable(&virge->bios_rom.mapping); 2.1387 + 2.1388 + mem_mapping_add(&virge->mmio_mapping, 0, 0, s3_virge_mmio_read, 2.1389 + s3_virge_mmio_read_w, 2.1390 + s3_virge_mmio_read_l, 2.1391 + s3_virge_mmio_write, 2.1392 + s3_virge_mmio_write_w, 2.1393 + s3_virge_mmio_write_l, 2.1394 + NULL, 2.1395 + 0, 2.1396 + virge); 2.1397 + mem_mapping_add(&virge->new_mmio_mapping, 0, 0, s3_virge_mmio_read, 2.1398 + s3_virge_mmio_read_w, 2.1399 + s3_virge_mmio_read_l, 2.1400 + s3_virge_mmio_write, 2.1401 + s3_virge_mmio_write_w, 2.1402 + s3_virge_mmio_write_l, 2.1403 + NULL, 2.1404 + 0, 2.1405 + virge); 2.1406 + mem_mapping_add(&virge->linear_mapping, 0, 0, svga_read_linear, 2.1407 + svga_readw_linear, 2.1408 + svga_readl_linear, 2.1409 + svga_write_linear, 2.1410 + svga_writew_linear, 2.1411 + svga_writel_linear, 2.1412 + NULL, 2.1413 + 0, 2.1414 + &virge->svga); 2.1415 + 2.1416 + io_sethandler(0x03c0, 0x0020, s3_virge_in, NULL, NULL, s3_virge_out, NULL, NULL, virge); 2.1417 + 2.1418 + virge->pci_regs[4] = 3; 2.1419 + virge->pci_regs[5] = 0; 2.1420 + virge->pci_regs[6] = 0; 2.1421 + virge->pci_regs[7] = 2; 2.1422 + virge->pci_regs[0x32] = 0x0c; 2.1423 + virge->pci_regs[0x3d] = 1; 2.1424 + virge->pci_regs[0x3e] = 4; 2.1425 + virge->pci_regs[0x3f] = 0xff; 2.1426 + 2.1427 + virge->virge_id_high = 0x8a; 2.1428 + virge->virge_id_low = 0x01; 2.1429 + virge->virge_rev = 0; 2.1430 + virge->virge_id = 0xe1; 2.1431 + 2.1432 + virge->svga.crtc[0x36] = 2 | (0 << 2) | (1 << 4); 2.1433 + virge->svga.crtc[0x37] = 1;// | (7 << 5); 2.1434 + virge->svga.crtc[0x53] = 1 << 3; 2.1435 + virge->svga.crtc[0x59] = 0x70; 2.1436 + 2.1437 + virge->svga.crtc[0x6c] = 0x01; 2.1438 + 2.1439 + virge->is_375 = 1; 2.1440 2.1441 pci_add(s3_virge_pci_read, s3_virge_pci_write, virge); 2.1442 2.1443 @@ -1521,6 +2654,9 @@ 2.1444 static void s3_virge_close(void *p) 2.1445 { 2.1446 virge_t *virge = (virge_t *)p; 2.1447 + FILE *f = fopen("vram.dmp", "wb"); 2.1448 + fwrite(virge->svga.vram, 4 << 20, 1, f); 2.1449 + fclose(f); 2.1450 2.1451 svga_close(&virge->svga); 2.1452 2.1453 @@ -1532,6 +2668,11 @@ 2.1454 return rom_present("roms/s3virge.bin"); 2.1455 } 2.1456 2.1457 +static int s3_virge_375_available() 2.1458 +{ 2.1459 + return rom_present("roms/86c375_1.bin"); 2.1460 +} 2.1461 + 2.1462 static void s3_virge_speed_changed(void *p) 2.1463 { 2.1464 virge_t *virge = (virge_t *)p; 2.1465 @@ -1549,13 +2690,21 @@ 2.1466 static int s3_virge_add_status_info(char *s, int max_len, void *p) 2.1467 { 2.1468 virge_t *virge = (virge_t *)p; 2.1469 + int cur_len; 2.1470 + char temps[256]; 2.1471 + 2.1472 + cur_len = svga_add_status_info(s, cur_len, &virge->svga); 2.1473 + sprintf(temps, "%f Mpixels/sec\n%f ktris/sec\n", (double)virge->pixel_count/1000000.0, (double)virge->tri_count/1000.0); 2.1474 + strncat(s, temps, cur_len); 2.1475 + cur_len -= strlen(temps); 2.1476 + virge->pixel_count = virge->tri_count = 0; 2.1477 2.1478 - return svga_add_status_info(s, max_len, &virge->svga); 2.1479 + return max_len - cur_len; 2.1480 } 2.1481 2.1482 device_t s3_virge_device = 2.1483 { 2.1484 - "Diamond Stealth 3D 2000 (S3 VIRGE)", 2.1485 + "Diamond Stealth 3D 2000 (S3 ViRGE)", 2.1486 DEVICE_NOT_WORKING, 2.1487 s3_virge_init, 2.1488 s3_virge_close, 2.1489 @@ -1564,3 +2713,15 @@ 2.1490 s3_virge_force_redraw, 2.1491 s3_virge_add_status_info 2.1492 }; 2.1493 + 2.1494 +device_t s3_virge_375_device = 2.1495 +{ 2.1496 + "S3 ViRGE/DX", 2.1497 + DEVICE_NOT_WORKING, 2.1498 + s3_virge_375_init, 2.1499 + s3_virge_close, 2.1500 + s3_virge_375_available, 2.1501 + s3_virge_speed_changed, 2.1502 + s3_virge_force_redraw, 2.1503 + s3_virge_add_status_info 2.1504 +};
3.1 --- a/src/vid_s3_virge.h Sun May 11 14:09:13 2014 +0100 3.2 +++ b/src/vid_s3_virge.h Mon May 26 18:13:23 2014 +0100 3.3 @@ -1,1 +1,2 @@ 3.4 extern device_t s3_virge_device; 3.5 +extern device_t s3_virge_375_device;
4.1 --- a/src/video.c Sun May 11 14:09:13 2014 +0100 4.2 +++ b/src/video.c Mon May 26 18:13:23 2014 +0100 4.3 @@ -54,6 +54,7 @@ 4.4 {"Paradise Bahamas 64 (S3 Vision864)", &s3_bahamas64_device, GFX_BAHAMAS64}, 4.5 {"Number Nine 9FX (S3 Trio64)", &s3_9fx_device, GFX_N9_9FX}, 4.6 {"Diamond Stealth 3D 2000 (S3 ViRGE)", &s3_virge_device, GFX_VIRGE}, 4.7 + {"S3 ViRGE/DX", &s3_virge_375_device, GFX_VIRGEDX}, 4.8 {"Trident TGUI9440", &tgui9440_device, GFX_TGUI9440}, 4.9 {"VGA", &vga_device, GFX_VGA}, 4.10 {"ATI VGA Edge-16 (ATI-18800)", &ati18800_device, GFX_VGAEDGE16},
5.1 --- a/src/win.c Sun May 11 14:09:13 2014 +0100 5.2 +++ b/src/win.c Mon May 26 18:13:23 2014 +0100 5.3 @@ -51,7 +51,7 @@ 5.4 #define TIMER_1SEC 1 5.5 5.6 int winsizex=640,winsizey=480; 5.7 -int gfx_present[18]; 5.8 +int gfx_present[19]; 5.9 int wakeups,wokeups; 5.10 #undef cs 5.11 CRITICAL_SECTION cs; 5.12 @@ -387,13 +387,13 @@ 5.13 } 5.14 5.15 5.16 - for (c = 0; c < 18; c++) 5.17 + for (c = 0; c < 19; c++) 5.18 gfx_present[c] = video_card_available(video_old_to_new(c)); 5.19 5.20 if (!video_card_available(video_old_to_new(gfxcard))) 5.21 { 5.22 if (romset!=-1) MessageBox(hwnd,"Configured video BIOS not available.\nDefaulting to available romset.","PCem error",MB_OK); 5.23 - for (c = 17; c >= 0; c--) 5.24 + for (c = 18; c >= 0; c--) 5.25 { 5.26 if (gfx_present[c]) 5.27 {