Commit | Line | Data |
---|---|---|
030d1121 | 1 | /*************************************************************************** |
2 | * Copyright (C) 2010 PCSX4ALL Team * | |
3 | * Copyright (C) 2010 Unai * | |
4 | * Copyright (C) 2016 Senquack (dansilsby <AT> gmail <DOT> com) * | |
5 | * * | |
6 | * This program is free software; you can redistribute it and/or modify * | |
7 | * it under the terms of the GNU General Public License as published by * | |
8 | * the Free Software Foundation; either version 2 of the License, or * | |
9 | * (at your option) any later version. * | |
10 | * * | |
11 | * This program is distributed in the hope that it will be useful, * | |
12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of * | |
13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * | |
14 | * GNU General Public License for more details. * | |
15 | * * | |
16 | * You should have received a copy of the GNU General Public License * | |
17 | * along with this program; if not, write to the * | |
18 | * Free Software Foundation, Inc., * | |
19 | * 51 Franklin Street, Fifth Floor, Boston, MA 02111-1307 USA. * | |
20 | ***************************************************************************/ | |
21 | ||
22 | #ifndef GPU_UNAI_H | |
23 | #define GPU_UNAI_H | |
24 | ||
25 | #include "gpu.h" | |
26 | ||
27 | // Header shared between both standalone gpu_unai (gpu.cpp) and new | |
28 | // gpulib-compatible gpu_unai (gpulib_if.cpp) | |
29 | // -> Anything here should be for gpu_unai's private use. <- | |
30 | ||
31 | /////////////////////////////////////////////////////////////////////////////// | |
32 | // Compile Options | |
33 | ||
34 | //#define ENABLE_GPU_NULL_SUPPORT // Enables NullGPU support | |
35 | //#define ENABLE_GPU_LOG_SUPPORT // Enables gpu logger, very slow only for windows debugging | |
36 | //#define ENABLE_GPU_ARMV7 // Enables ARMv7 optimized assembly | |
37 | ||
38 | //Poly routine options (default is integer math and accurate division) | |
39 | //#define GPU_UNAI_USE_FLOATMATH // Use float math in poly routines | |
40 | //#define GPU_UNAI_USE_FLOAT_DIV_MULTINV // If GPU_UNAI_USE_FLOATMATH is defined, | |
41 | // use multiply-by-inverse for division | |
42 | //#define GPU_UNAI_USE_INT_DIV_MULTINV // If GPU_UNAI_USE_FLOATMATH is *not* | |
43 | // defined, use old inaccurate division | |
44 | ||
45 | ||
46 | #define GPU_INLINE static inline __attribute__((always_inline)) | |
47 | #define INLINE static inline __attribute__((always_inline)) | |
48 | ||
49 | #define u8 uint8_t | |
50 | #define s8 int8_t | |
51 | #define u16 uint16_t | |
52 | #define s16 int16_t | |
53 | #define u32 uint32_t | |
54 | #define s32 int32_t | |
55 | #define s64 int64_t | |
ebcd59b7 PC |
56 | #define u64 uint64_t |
57 | ||
58 | typedef union { | |
59 | struct { | |
60 | u16 r, g, b; | |
61 | } c; | |
62 | u64 raw; | |
63 | } gcol_t; | |
030d1121 | 64 | |
4949d4ff PC |
65 | typedef struct { |
66 | u32 v; | |
67 | } le32_t; | |
68 | ||
69 | typedef struct { | |
70 | u16 v; | |
71 | } le16_t; | |
72 | ||
73 | static inline u32 le32_to_u32(le32_t le) | |
74 | { | |
75 | return LE32TOH(le.v); | |
76 | } | |
77 | ||
78 | static inline s32 le32_to_s32(le32_t le) | |
79 | { | |
80 | return (int32_t) LE32TOH(le.v); | |
81 | } | |
82 | ||
83 | static inline u32 le32_raw(le32_t le) | |
84 | { | |
85 | return le.v; | |
86 | } | |
87 | ||
88 | static inline le32_t u32_to_le32(u32 u) | |
89 | { | |
90 | return (le32_t){ .v = HTOLE32(u) }; | |
91 | } | |
92 | ||
93 | static inline u16 le16_to_u16(le16_t le) | |
94 | { | |
95 | return LE16TOH(le.v); | |
96 | } | |
97 | ||
98 | static inline s16 le16_to_s16(le16_t le) | |
99 | { | |
100 | return (int16_t) LE16TOH(le.v); | |
101 | } | |
102 | ||
103 | static inline u16 le16_raw(le16_t le) | |
104 | { | |
105 | return le.v; | |
106 | } | |
107 | ||
108 | static inline le16_t u16_to_le16(u16 u) | |
109 | { | |
110 | return (le16_t){ .v = HTOLE16(u) }; | |
111 | } | |
112 | ||
030d1121 | 113 | union PtrUnion |
114 | { | |
4949d4ff PC |
115 | le32_t *U4; |
116 | le16_t *U2; | |
030d1121 | 117 | u8 *U1; |
030d1121 | 118 | void *ptr; |
119 | }; | |
120 | ||
121 | union GPUPacket | |
122 | { | |
4949d4ff PC |
123 | le32_t U4[16]; |
124 | le16_t U2[32]; | |
030d1121 | 125 | u8 U1[64]; |
030d1121 | 126 | }; |
127 | ||
128 | template<class T> static inline void SwapValues(T &x, T &y) | |
129 | { | |
130 | T tmp(x); x = y; y = tmp; | |
131 | } | |
132 | ||
133 | template<typename T> | |
134 | static inline T Min2 (const T a, const T b) | |
135 | { | |
136 | return (a<b)?a:b; | |
137 | } | |
138 | ||
139 | template<typename T> | |
140 | static inline T Min3 (const T a, const T b, const T c) | |
141 | { | |
142 | return Min2(Min2(a,b),c); | |
143 | } | |
144 | ||
145 | template<typename T> | |
146 | static inline T Max2 (const T a, const T b) | |
147 | { | |
148 | return (a>b)?a:b; | |
149 | } | |
150 | ||
151 | template<typename T> | |
152 | static inline T Max3 (const T a, const T b, const T c) | |
153 | { | |
154 | return Max2(Max2(a,b),c); | |
155 | } | |
156 | ||
157 | ||
158 | /////////////////////////////////////////////////////////////////////////////// | |
159 | // GPU Raster Macros | |
160 | ||
161 | // Convert 24bpp color parameter of GPU command to 16bpp (15bpp + mask bit) | |
162 | #define GPU_RGB16(rgb) ((((rgb)&0xF80000)>>9)|(((rgb)&0xF800)>>6)|(((rgb)&0xF8)>>3)) | |
163 | ||
164 | // Sign-extend 11-bit coordinate command param | |
165 | #define GPU_EXPANDSIGN(x) (((s32)(x)<<(32-11))>>(32-11)) | |
166 | ||
167 | // Max difference between any two X or Y primitive coordinates | |
168 | #define CHKMAX_X 1024 | |
169 | #define CHKMAX_Y 512 | |
170 | ||
171 | #define FRAME_BUFFER_SIZE (1024*512*2) | |
172 | #define FRAME_WIDTH 1024 | |
173 | #define FRAME_HEIGHT 512 | |
174 | #define FRAME_OFFSET(x,y) (((y)<<10)+(x)) | |
175 | #define FRAME_BYTE_STRIDE 2048 | |
176 | #define FRAME_BYTES_PER_PIXEL 2 | |
177 | ||
178 | static inline s32 GPU_DIV(s32 rs, s32 rt) | |
179 | { | |
180 | return rt ? (rs / rt) : (0); | |
181 | } | |
182 | ||
183 | // 'Unsafe' version of above that doesn't check for div-by-zero | |
184 | #define GPU_FAST_DIV(rs, rt) ((signed)(rs) / (signed)(rt)) | |
185 | ||
186 | struct gpu_unai_t { | |
187 | u32 GPU_GP1; | |
188 | GPUPacket PacketBuffer; | |
4949d4ff | 189 | le16_t *vram; |
030d1121 | 190 | |
43047988 | 191 | #ifdef USE_GPULIB |
8999ff50 | 192 | le16_t *downscale_vram; |
43047988 | 193 | #endif |
030d1121 | 194 | //////////////////////////////////////////////////////////////////////////// |
195 | // Variables used only by older standalone version of gpu_unai (gpu.cpp) | |
196 | #ifndef USE_GPULIB | |
197 | u32 GPU_GP0; | |
198 | u32 tex_window; // Current texture window vals (set by GP0(E2h) cmd) | |
199 | s32 PacketCount; | |
200 | s32 PacketIndex; | |
201 | bool fb_dirty; // Framebuffer is dirty (according to GPU) | |
202 | ||
203 | // Display status | |
204 | // NOTE: Standalone older gpu_unai didn't care about horiz display range | |
205 | u16 DisplayArea[6]; // [0] : Start of display area (in VRAM) X | |
206 | // [1] : Start of display area (in VRAM) Y | |
207 | // [2] : Display mode resolution HORIZONTAL | |
208 | // [3] : Display mode resolution VERTICAL | |
209 | // [4] : Vertical display range (on TV) START | |
210 | // [5] : Vertical display range (on TV) END | |
211 | ||
212 | //////////////////////////////////////////////////////////////////////////// | |
213 | // Dma Transfers info | |
214 | struct { | |
215 | s32 px,py; | |
216 | s32 x_end,y_end; | |
4949d4ff | 217 | le16_t* pvram; |
030d1121 | 218 | u32 *last_dma; // Last dma pointer |
219 | bool FrameToRead; // Load image in progress | |
220 | bool FrameToWrite; // Store image in progress | |
221 | } dma; | |
222 | ||
223 | //////////////////////////////////////////////////////////////////////////// | |
224 | // Frameskip | |
225 | struct { | |
226 | int skipCount; // Frame skip (0,1,2,3...) | |
227 | bool isSkip; // Skip frame (according to GPU) | |
228 | bool skipFrame; // Skip this frame (according to frame skip) | |
229 | bool wasSkip; // Skip frame old value (according to GPU) | |
230 | bool skipGPU; // Skip GPU primitives | |
231 | } frameskip; | |
232 | #endif | |
233 | // END of standalone gpu_unai variables | |
234 | //////////////////////////////////////////////////////////////////////////// | |
235 | ||
236 | u32 TextureWindowCur; // Current setting from last GP0(0xE2) cmd (raw form) | |
237 | u8 TextureWindow[4]; // [0] : Texture window offset X | |
238 | // [1] : Texture window offset Y | |
239 | // [2] : Texture window mask X | |
240 | // [3] : Texture window mask Y | |
241 | ||
242 | u16 DrawingArea[4]; // [0] : Drawing area top left X | |
243 | // [1] : Drawing area top left Y | |
244 | // [2] : Drawing area bottom right X | |
245 | // [3] : Drawing area bottom right Y | |
246 | ||
247 | s16 DrawingOffset[2]; // [0] : Drawing offset X (signed) | |
248 | // [1] : Drawing offset Y (signed) | |
249 | ||
4949d4ff PC |
250 | le16_t* TBA; // Ptr to current texture in VRAM |
251 | le16_t* CBA; // Ptr to current CLUT in VRAM | |
030d1121 | 252 | |
253 | //////////////////////////////////////////////////////////////////////////// | |
254 | // Inner Loop parameters | |
255 | ||
256 | // 22.10 Fixed-pt texture coords, mask, scanline advance | |
257 | // NOTE: U,V are no longer packed together into one u32, this proved to be | |
258 | // too imprecise, leading to pixel dropouts. Example: NFS3's skybox. | |
259 | u32 u, v; | |
260 | u32 u_msk, v_msk; | |
261 | s32 u_inc, v_inc; | |
262 | ||
263 | // Color for Gouraud-shaded prims | |
ebcd59b7 | 264 | // Fixed-pt 8.8 rgb triplet |
030d1121 | 265 | // Packed fixed-pt 8.3:8.3:8.2 rgb triplet |
ebcd59b7 PC |
266 | // layout: ccccccccXXXXXXXX for c in [r, g, b] |
267 | // ^ bit 16 | |
268 | gcol_t gCol; | |
269 | gcol_t gInc; // Increment along scanline for gCol | |
030d1121 | 270 | |
271 | // Color for flat-shaded, texture-blended prims | |
272 | u8 r5, g5, b5; // 5-bit light for undithered prims | |
273 | u8 r8, g8, b8; // 8-bit light for dithered prims | |
274 | ||
275 | // Color for flat-shaded, untextured prims | |
276 | u16 PixelData; // bgr555 color for untextured flat-shaded polys | |
277 | ||
278 | // End of inner Loop parameters | |
279 | //////////////////////////////////////////////////////////////////////////// | |
280 | ||
281 | ||
282 | u8 blit_mask; // Determines what pixels to skip when rendering. | |
283 | // Only useful on low-resolution devices using | |
284 | // a simple pixel-dropping downscaler for PS1 | |
285 | // high-res modes. See 'pixel_skip' option. | |
286 | ||
287 | u8 ilace_mask; // Determines what lines to skip when rendering. | |
288 | // Normally 0 when PS1 240 vertical res is in | |
289 | // use and ilace_force is 0. When running in | |
290 | // PS1 480 vertical res on a low-resolution | |
291 | // device (320x240), will usually be set to 1 | |
292 | // so odd lines are not rendered. (Unless future | |
293 | // full-screen scaling option is in use ..TODO) | |
294 | ||
295 | bool prog_ilace_flag; // Tracks successive frames for 'prog_ilace' option | |
296 | ||
297 | u8 BLEND_MODE; | |
298 | u8 TEXT_MODE; | |
299 | u8 Masking; | |
300 | ||
301 | u16 PixelMSB; | |
302 | ||
303 | gpu_unai_config_t config; | |
304 | ||
305 | u8 LightLUT[32*32]; // 5-bit lighting LUT (gpu_inner_light.h) | |
306 | u32 DitherMatrix[64]; // Matrix of dither coefficients | |
307 | }; | |
308 | ||
309 | static gpu_unai_t gpu_unai; | |
310 | ||
311 | // Global config that frontend can alter.. Values are read in GPU_init(). | |
312 | // TODO: if frontend menu modifies a setting, add a function that can notify | |
313 | // GPU plugin to use new setting. | |
314 | gpu_unai_config_t gpu_unai_config_ext; | |
315 | ||
316 | /////////////////////////////////////////////////////////////////////////////// | |
317 | // Internal inline funcs to get option status: (Allows flexibility) | |
318 | static inline bool LightingEnabled() | |
319 | { | |
320 | return gpu_unai.config.lighting; | |
321 | } | |
322 | ||
323 | static inline bool FastLightingEnabled() | |
324 | { | |
325 | return gpu_unai.config.fast_lighting; | |
326 | } | |
327 | ||
328 | static inline bool BlendingEnabled() | |
329 | { | |
330 | return gpu_unai.config.blending; | |
331 | } | |
332 | ||
333 | static inline bool DitheringEnabled() | |
334 | { | |
335 | return gpu_unai.config.dithering; | |
336 | } | |
337 | ||
338 | // For now, this is just for development/experimentation purposes.. | |
339 | // If modified to return true, it will allow ignoring the status register | |
340 | // bit 9 setting (dither enable). It will still restrict dithering only | |
341 | // to Gouraud-shaded or texture-blended polys. | |
342 | static inline bool ForcedDitheringEnabled() | |
343 | { | |
344 | return false; | |
345 | } | |
346 | ||
347 | static inline bool ProgressiveInterlaceEnabled() | |
348 | { | |
349 | #ifdef USE_GPULIB | |
350 | // Using this old option greatly decreases quality of image. Disabled | |
351 | // for now when using new gpulib, since it also adds more work in loops. | |
352 | return false; | |
353 | #else | |
354 | return gpu_unai.config.prog_ilace; | |
355 | #endif | |
356 | } | |
357 | ||
358 | // For now, 320x240 output resolution is assumed, using simple line-skipping | |
359 | // and pixel-skipping downscaler. | |
360 | // TODO: Flesh these out so they return useful values based on whether | |
361 | // running on higher-res device or a resampling downscaler is enabled. | |
362 | static inline bool PixelSkipEnabled() | |
363 | { | |
43047988 | 364 | return gpu_unai.config.pixel_skip || gpu_unai.config.scale_hires; |
030d1121 | 365 | } |
366 | ||
367 | static inline bool LineSkipEnabled() | |
368 | { | |
369 | return true; | |
370 | } | |
371 | ||
372 | #endif // GPU_UNAI_H |