1 /***************************************************************************
2 * Copyright (C) 2010 PCSX4ALL Team *
3 * Copyright (C) 2010 Unai *
4 * Copyright (C) 2016 Senquack (dansilsby <AT> gmail <DOT> com) *
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. *
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. *
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 ***************************************************************************/
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. <-
31 ///////////////////////////////////////////////////////////////////////////////
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
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
46 #define GPU_INLINE static inline __attribute__((always_inline))
47 #define INLINE static inline __attribute__((always_inline))
73 static inline u32 le32_to_u32(le32_t le)
78 static inline s32 le32_to_s32(le32_t le)
80 return (int32_t) LE32TOH(le.v);
83 static inline u32 le32_raw(le32_t le)
88 static inline le32_t u32_to_le32(u32 u)
90 return (le32_t){ .v = HTOLE32(u) };
93 static inline u16 le16_to_u16(le16_t le)
98 static inline s16 le16_to_s16(le16_t le)
100 return (int16_t) LE16TOH(le.v);
103 static inline u16 le16_raw(le16_t le)
108 static inline le16_t u16_to_le16(u16 u)
110 return (le16_t){ .v = HTOLE16(u) };
128 template<class T> static inline void SwapValues(T &x, T &y)
130 T tmp(x); x = y; y = tmp;
134 static inline T Min2 (const T a, const T b)
140 static inline T Min3 (const T a, const T b, const T c)
142 return Min2(Min2(a,b),c);
146 static inline T Max2 (const T a, const T b)
152 static inline T Max3 (const T a, const T b, const T c)
154 return Max2(Max2(a,b),c);
158 ///////////////////////////////////////////////////////////////////////////////
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))
164 // Sign-extend 11-bit coordinate command param
165 #define GPU_EXPANDSIGN(x) (((s32)(x)<<(32-11))>>(32-11))
167 // Max difference between any two X or Y primitive coordinates
168 #define CHKMAX_X 1024
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
178 static inline s32 GPU_DIV(s32 rs, s32 rt)
180 return rt ? (rs / rt) : (0);
183 // 'Unsafe' version of above that doesn't check for div-by-zero
184 #define GPU_FAST_DIV(rs, rt) ((signed)(rs) / (signed)(rt))
188 GPUPacket PacketBuffer;
192 le16_t *downscale_vram;
194 ////////////////////////////////////////////////////////////////////////////
195 // Variables used only by older standalone version of gpu_unai (gpu.cpp)
198 u32 tex_window; // Current texture window vals (set by GP0(E2h) cmd)
201 bool fb_dirty; // Framebuffer is dirty (according to GPU)
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
212 ////////////////////////////////////////////////////////////////////////////
213 // Dma Transfers info
218 u32 *last_dma; // Last dma pointer
219 bool FrameToRead; // Load image in progress
220 bool FrameToWrite; // Store image in progress
223 ////////////////////////////////////////////////////////////////////////////
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
233 // END of standalone gpu_unai variables
234 ////////////////////////////////////////////////////////////////////////////
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
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
247 s16 DrawingOffset[2]; // [0] : Drawing offset X (signed)
248 // [1] : Drawing offset Y (signed)
250 le16_t* TBA; // Ptr to current texture in VRAM
251 le16_t* CBA; // Ptr to current CLUT in VRAM
253 ////////////////////////////////////////////////////////////////////////////
254 // Inner Loop parameters
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.
263 // Color for Gouraud-shaded prims
264 // Fixed-pt 8.8 rgb triplet
265 // Packed fixed-pt 8.3:8.3:8.2 rgb triplet
266 // layout: ccccccccXXXXXXXX for c in [r, g, b]
269 gcol_t gInc; // Increment along scanline for gCol
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
275 // Color for flat-shaded, untextured prims
276 u16 PixelData; // bgr555 color for untextured flat-shaded polys
278 // End of inner Loop parameters
279 ////////////////////////////////////////////////////////////////////////////
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.
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)
295 bool prog_ilace_flag; // Tracks successive frames for 'prog_ilace' option
303 gpu_unai_config_t config;
305 u8 LightLUT[32*32]; // 5-bit lighting LUT (gpu_inner_light.h)
306 u32 DitherMatrix[64]; // Matrix of dither coefficients
309 static gpu_unai_t gpu_unai;
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;
316 ///////////////////////////////////////////////////////////////////////////////
317 // Internal inline funcs to get option status: (Allows flexibility)
318 static inline bool LightingEnabled()
320 return gpu_unai.config.lighting;
323 static inline bool FastLightingEnabled()
325 return gpu_unai.config.fast_lighting;
328 static inline bool BlendingEnabled()
330 return gpu_unai.config.blending;
333 static inline bool DitheringEnabled()
335 return gpu_unai.config.dithering;
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()
347 static inline bool ProgressiveInterlaceEnabled()
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.
354 return gpu_unai.config.prog_ilace;
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()
364 return gpu_unai.config.pixel_skip || gpu_unai.config.scale_hires;
367 static inline bool LineSkipEnabled()