0bfe8d59 |
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_senquack (gpu.cpp) and new |
28 | // gpulib-compatible gpu_senquack (gpulib_if.cpp) |
29 | // -> Anything here should be for gpu_senquack'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 |
56 | |
57 | union PtrUnion |
58 | { |
59 | u32 *U4; |
60 | s32 *S4; |
61 | u16 *U2; |
62 | s16 *S2; |
63 | u8 *U1; |
64 | s8 *S1; |
65 | void *ptr; |
66 | }; |
67 | |
68 | union GPUPacket |
69 | { |
70 | u32 U4[16]; |
71 | s32 S4[16]; |
72 | u16 U2[32]; |
73 | s16 S2[32]; |
74 | u8 U1[64]; |
75 | s8 S1[64]; |
76 | }; |
77 | |
78 | template<class T> static inline void SwapValues(T &x, T &y) |
79 | { |
80 | T tmp(x); x = y; y = tmp; |
81 | } |
82 | |
83 | template<typename T> |
84 | static inline T Min2 (const T a, const T b) |
85 | { |
86 | return (a<b)?a:b; |
87 | } |
88 | |
89 | template<typename T> |
90 | static inline T Min3 (const T a, const T b, const T c) |
91 | { |
92 | return Min2(Min2(a,b),c); |
93 | } |
94 | |
95 | template<typename T> |
96 | static inline T Max2 (const T a, const T b) |
97 | { |
98 | return (a>b)?a:b; |
99 | } |
100 | |
101 | template<typename T> |
102 | static inline T Max3 (const T a, const T b, const T c) |
103 | { |
104 | return Max2(Max2(a,b),c); |
105 | } |
106 | |
107 | |
108 | /////////////////////////////////////////////////////////////////////////////// |
109 | // GPU Raster Macros |
110 | |
111 | // Convert 24bpp color parameter of GPU command to 16bpp (15bpp + mask bit) |
112 | #define GPU_RGB16(rgb) ((((rgb)&0xF80000)>>9)|(((rgb)&0xF800)>>6)|(((rgb)&0xF8)>>3)) |
113 | |
114 | // Sign-extend 11-bit coordinate command param |
115 | #define GPU_EXPANDSIGN(x) (((s32)(x)<<(32-11))>>(32-11)) |
116 | |
117 | // Max difference between any two X or Y primitive coordinates |
118 | #define CHKMAX_X 1024 |
119 | #define CHKMAX_Y 512 |
120 | |
121 | #define FRAME_BUFFER_SIZE (1024*512*2) |
122 | #define FRAME_WIDTH 1024 |
123 | #define FRAME_HEIGHT 512 |
124 | #define FRAME_OFFSET(x,y) (((y)<<10)+(x)) |
125 | #define FRAME_BYTE_STRIDE 2048 |
126 | #define FRAME_BYTES_PER_PIXEL 2 |
127 | |
128 | static inline s32 GPU_DIV(s32 rs, s32 rt) |
129 | { |
130 | return rt ? (rs / rt) : (0); |
131 | } |
132 | |
133 | // 'Unsafe' version of above that doesn't check for div-by-zero |
134 | #define GPU_FAST_DIV(rs, rt) ((signed)(rs) / (signed)(rt)) |
135 | |
136 | struct gpu_senquack_t { |
137 | u32 GPU_GP1; |
138 | GPUPacket PacketBuffer; |
139 | u16 *vram; |
140 | |
141 | #ifndef USE_GPULIB |
142 | u32 GPU_GP0; |
143 | u32 tex_window; // Current texture window vals (set by GP0(E2h) cmd) |
144 | s32 PacketCount; |
145 | s32 PacketIndex; |
146 | bool fb_dirty; // Framebuffer is dirty (according to GPU) |
147 | |
148 | // Display status |
149 | // NOTE: Standalone older gpu_senquack didn't care about horiz display range |
150 | u16 DisplayArea[6]; // [0] : Start of display area (in VRAM) X |
151 | // [1] : Start of display area (in VRAM) Y |
152 | // [2] : Display mode resolution HORIZONTAL |
153 | // [3] : Display mode resolution VERTICAL |
154 | // [4] : Vertical display range (on TV) START |
155 | // [5] : Vertical display range (on TV) END |
156 | |
157 | //////////////////////////////////////////////////////////////////////////// |
158 | // Dma Transfers info |
159 | struct { |
160 | s32 px,py; |
161 | s32 x_end,y_end; |
162 | u16* pvram; |
163 | u32 *last_dma; // Last dma pointer |
164 | bool FrameToRead; // Load image in progress |
165 | bool FrameToWrite; // Store image in progress |
166 | } dma; |
167 | |
168 | //////////////////////////////////////////////////////////////////////////// |
169 | // Frameskip |
170 | struct { |
171 | int skipCount; // Frame skip (0,1,2,3...) |
172 | bool isSkip; // Skip frame (according to GPU) |
173 | bool skipFrame; // Skip this frame (according to frame skip) |
174 | bool wasSkip; // Skip frame old value (according to GPU) |
175 | bool skipGPU; // Skip GPU primitives |
176 | } frameskip; |
177 | #endif |
178 | // END of standalone gpu_senquack variables |
179 | //////////////////////////////////////////////////////////////////////////// |
180 | |
181 | u32 TextureWindowCur; // Current setting from last GP0(0xE2) cmd (raw form) |
182 | u8 TextureWindow[4]; // [0] : Texture window offset X |
183 | // [1] : Texture window offset Y |
184 | // [2] : Texture window mask X |
185 | // [3] : Texture window mask Y |
186 | |
187 | u16 DrawingArea[4]; // [0] : Drawing area top left X |
188 | // [1] : Drawing area top left Y |
189 | // [2] : Drawing area bottom right X |
190 | // [3] : Drawing area bottom right Y |
191 | |
192 | s16 DrawingOffset[2]; // [0] : Drawing offset X (signed) |
193 | // [1] : Drawing offset Y (signed) |
194 | |
195 | u16* TBA; // Ptr to current texture in VRAM |
196 | u16* CBA; // Ptr to current CLUT in VRAM |
197 | |
198 | //////////////////////////////////////////////////////////////////////////// |
199 | // Inner Loop parameters |
200 | |
201 | // 22.10 Fixed-pt texture coords, mask, scanline advance |
202 | // NOTE: U,V are no longer packed together into one u32, this proved to be |
203 | // too imprecise, leading to pixel dropouts. Example: NFS3's skybox. |
204 | u32 u, v; |
205 | u32 u_msk, v_msk; |
206 | s32 u_inc, v_inc; |
207 | |
208 | // Color for Gouraud-shaded prims |
209 | // Packed fixed-pt 8.3:8.3:8.2 rgb triplet |
210 | // layout: rrrrrrrrXXXggggggggXXXbbbbbbbbXX |
211 | // ^ bit 31 ^ bit 0 |
212 | u32 gCol; |
213 | u32 gInc; // Increment along scanline for gCol |
214 | |
215 | // Color for flat-shaded, texture-blended prims |
216 | u8 r5, g5, b5; // 5-bit light for undithered prims |
217 | u8 r8, g8, b8; // 8-bit light for dithered prims |
218 | |
219 | // Color for flat-shaded, untextured prims |
220 | u16 PixelData; // bgr555 color for untextured flat-shaded polys |
221 | |
222 | // End of inner Loop parameters |
223 | //////////////////////////////////////////////////////////////////////////// |
224 | |
225 | |
226 | u8 blit_mask; // Determines what pixels to skip when rendering. |
227 | // Only useful on low-resolution devices using |
228 | // a simple pixel-dropping downscaler for PS1 |
229 | // high-res modes. See 'pixel_skip' option. |
230 | |
231 | u8 ilace_mask; // Determines what lines to skip when rendering. |
232 | // Normally 0 when PS1 240 vertical res is in |
233 | // use and ilace_force is 0. When running in |
234 | // PS1 480 vertical res on a low-resolution |
235 | // device (320x240), will usually be set to 1 |
236 | // so odd lines are not rendered. (Unless future |
237 | // full-screen scaling option is in use ..TODO) |
238 | |
239 | bool prog_ilace_flag; // Tracks successive frames for 'prog_ilace' option |
240 | |
241 | u8 BLEND_MODE; |
242 | u8 TEXT_MODE; |
243 | u8 Masking; |
244 | |
245 | u16 PixelMSB; |
246 | |
247 | gpu_senquack_config_t config; |
248 | |
249 | u8 LightLUT[32*32]; // 5-bit lighting LUT (gpu_inner_light.h) |
250 | u32 DitherMatrix[64]; // Matrix of dither coefficients |
251 | }; |
252 | |
253 | static gpu_senquack_t gpu_senquack; |
254 | |
255 | // Global config that frontend can alter.. Values are read in GPU_init(). |
256 | // TODO: if frontend menu modifies a setting, add a function that can notify |
257 | // GPU plugin to use new setting. |
258 | gpu_senquack_config_t gpu_senquack_config_ext; |
259 | |
260 | /////////////////////////////////////////////////////////////////////////////// |
261 | // Internal inline funcs to get option status: (Allows flexibility) |
262 | static inline bool LightingEnabled() |
263 | { |
264 | return gpu_senquack.config.lighting; |
265 | } |
266 | |
267 | static inline bool FastLightingEnabled() |
268 | { |
269 | return gpu_senquack.config.fast_lighting; |
270 | } |
271 | |
272 | static inline bool BlendingEnabled() |
273 | { |
274 | return gpu_senquack.config.blending; |
275 | } |
276 | |
277 | static inline bool DitheringEnabled() |
278 | { |
279 | return gpu_senquack.config.dithering; |
280 | } |
281 | |
282 | // For now, this is just for development/experimentation purposes.. |
283 | // If modified to return true, it will allow ignoring the status register |
284 | // bit 9 setting (dither enable). It will still restrict dithering only |
285 | // to Gouraud-shaded or texture-blended polys. |
286 | static inline bool ForcedDitheringEnabled() |
287 | { |
288 | return false; |
289 | } |
290 | |
291 | static inline bool ProgressiveInterlaceEnabled() |
292 | { |
293 | #ifdef USE_GPULIB |
294 | // Using this old option greatly decreases quality of image. Disabled |
295 | // for now when using new gpulib, since it also adds more work in loops. |
296 | return false; |
297 | #else |
298 | return gpu_senquack.config.prog_ilace; |
299 | #endif |
300 | } |
301 | |
302 | // For now, 320x240 output resolution is assumed, using simple line-skipping |
303 | // and pixel-skipping downscaler. |
304 | // TODO: Flesh these out so they return useful values based on whether |
305 | // running on higher-res device or a resampling downscaler is enabled. |
306 | static inline bool PixelSkipEnabled() |
307 | { |
308 | return gpu_senquack.config.pixel_skip || gpu_senquack.config.scale_hires; |
309 | } |
310 | |
311 | static inline bool LineSkipEnabled() |
312 | { |
313 | return true; |
314 | } |
315 | |
316 | #endif // GPU_UNAI_H |