dolphin/Source/Core/VideoCommon/XFMemory.h

312 lines
6.7 KiB
C
Raw Normal View History

// Copyright 2008 Dolphin Emulator Project
2015-05-17 17:08:10 -06:00
// Licensed under GPLv2+
// Refer to the license.txt file included.
#pragma once
#include "Common/CommonTypes.h"
#include "VideoCommon/CPMemory.h"
class DataReader;
// Lighting
// Projection
enum : u32
{
XF_TEXPROJ_ST = 0,
XF_TEXPROJ_STQ = 1
};
// Input form
enum : u32
{
XF_TEXINPUT_AB11 = 0,
XF_TEXINPUT_ABC1 = 1
};
// Texture generation type
enum : u32
{
XF_TEXGEN_REGULAR = 0,
XF_TEXGEN_EMBOSS_MAP = 1, // Used when bump mapping
XF_TEXGEN_COLOR_STRGBC0 = 2,
XF_TEXGEN_COLOR_STRGBC1 = 3
};
// Source row
enum : u32
{
XF_SRCGEOM_INROW = 0, // Input is abc
XF_SRCNORMAL_INROW = 1, // Input is abc
XF_SRCCOLORS_INROW = 2,
XF_SRCBINORMAL_T_INROW = 3, // Input is abc
XF_SRCBINORMAL_B_INROW = 4, // Input is abc
XF_SRCTEX0_INROW = 5,
XF_SRCTEX1_INROW = 6,
XF_SRCTEX2_INROW = 7,
XF_SRCTEX3_INROW = 8,
XF_SRCTEX4_INROW = 9,
XF_SRCTEX5_INROW = 10,
XF_SRCTEX6_INROW = 11,
XF_SRCTEX7_INROW = 12
};
// Control source
enum : u32
{
GX_SRC_REG = 0,
GX_SRC_VTX = 1
};
// Light diffuse attenuation function
enum : u32
{
LIGHTDIF_NONE = 0,
LIGHTDIF_SIGN = 1,
LIGHTDIF_CLAMP = 2
};
// Light attenuation function
enum : u32
{
LIGHTATTN_NONE = 0, // No attenuation
LIGHTATTN_SPEC = 1, // Point light attenuation
LIGHTATTN_DIR = 2, // Directional light attenuation
LIGHTATTN_SPOT = 3 // Spot light attenuation
};
// Projection type
enum : u32
{
GX_PERSPECTIVE = 0,
GX_ORTHOGRAPHIC = 1
};
// Registers and register ranges
enum
{
XFMEM_POSMATRICES = 0x000,
XFMEM_POSMATRICES_END = 0x100,
XFMEM_NORMALMATRICES = 0x400,
XFMEM_NORMALMATRICES_END = 0x460,
XFMEM_POSTMATRICES = 0x500,
XFMEM_POSTMATRICES_END = 0x600,
XFMEM_LIGHTS = 0x600,
XFMEM_LIGHTS_END = 0x680,
XFMEM_ERROR = 0x1000,
XFMEM_DIAG = 0x1001,
XFMEM_STATE0 = 0x1002,
XFMEM_STATE1 = 0x1003,
XFMEM_CLOCK = 0x1004,
XFMEM_CLIPDISABLE = 0x1005,
XFMEM_SETGPMETRIC = 0x1006,
XFMEM_VTXSPECS = 0x1008,
XFMEM_SETNUMCHAN = 0x1009,
XFMEM_SETCHAN0_AMBCOLOR = 0x100a,
XFMEM_SETCHAN1_AMBCOLOR = 0x100b,
XFMEM_SETCHAN0_MATCOLOR = 0x100c,
XFMEM_SETCHAN1_MATCOLOR = 0x100d,
XFMEM_SETCHAN0_COLOR = 0x100e,
XFMEM_SETCHAN1_COLOR = 0x100f,
XFMEM_SETCHAN0_ALPHA = 0x1010,
XFMEM_SETCHAN1_ALPHA = 0x1011,
XFMEM_DUALTEX = 0x1012,
XFMEM_SETMATRIXINDA = 0x1018,
XFMEM_SETMATRIXINDB = 0x1019,
XFMEM_SETVIEWPORT = 0x101a,
XFMEM_SETZSCALE = 0x101c,
XFMEM_SETZOFFSET = 0x101f,
XFMEM_SETPROJECTION = 0x1020,
// XFMEM_SETPROJECTIONB = 0x1021,
// XFMEM_SETPROJECTIONC = 0x1022,
// XFMEM_SETPROJECTIOND = 0x1023,
// XFMEM_SETPROJECTIONE = 0x1024,
// XFMEM_SETPROJECTIONF = 0x1025,
// XFMEM_SETPROJECTIONORTHO1 = 0x1026,
// XFMEM_SETPROJECTIONORTHO2 = 0x1027,
XFMEM_SETNUMTEXGENS = 0x103f,
XFMEM_SETTEXMTXINFO = 0x1040,
XFMEM_SETPOSMTXINFO = 0x1050,
};
union LitChannel
{
struct
{
u32 matsource : 1;
u32 enablelighting : 1;
u32 lightMask0_3 : 4;
u32 ambsource : 1;
u32 diffusefunc : 2; // LIGHTDIF_X
u32 attnfunc : 2; // LIGHTATTN_X
u32 lightMask4_7 : 4;
};
struct
{
u32 hex : 15;
u32 unused : 17;
};
struct
{
u32 dummy0 : 7;
u32 lightparams : 4;
u32 dummy1 : 21;
};
unsigned int GetFullLightMask() const
{
return enablelighting ? (lightMask0_3 | (lightMask4_7 << 4)) : 0;
}
};
union INVTXSPEC
{
struct
{
u32 numcolors : 2;
u32 numnormals : 2; // 0 - nothing, 1 - just normal, 2 - normals and binormals
u32 numtextures : 4;
u32 unused : 24;
};
u32 hex;
};
union TexMtxInfo
{
struct
{
u32 unknown : 1;
u32 projection : 1; // XF_TEXPROJ_X
u32 inputform : 1; // XF_TEXINPUT_X
u32 unknown2 : 1;
u32 texgentype : 3; // XF_TEXGEN_X
u32 sourcerow : 5; // XF_SRCGEOM_X
u32 embosssourceshift : 3; // what generated texcoord to use
u32 embosslightshift : 3; // light index that is used
};
u32 hex;
};
union PostMtxInfo
{
struct
{
u32 index : 6; // base row of dual transform matrix
u32 unused : 2;
u32 normalize : 1; // normalize before send operation
};
u32 hex;
};
union NumColorChannel
{
struct
{
u32 numColorChans : 2;
};
u32 hex;
};
union NumTexGen
{
struct
{
u32 numTexGens : 4;
};
u32 hex;
};
union DualTexInfo
{
struct
{
u32 enabled : 1;
};
u32 hex;
};
struct Light
{
u32 useless[3];
u8 color[4];
float cosatt[3]; // cos attenuation
float distatt[3]; // dist attenuation
union
{
struct
{
float dpos[3];
float ddir[3]; // specular lights only
};
struct
{
float sdir[3];
float shalfangle[3]; // specular lights only
};
};
};
struct Viewport
{
float wd;
float ht;
float zRange;
float xOrig;
float yOrig;
float farZ;
};
struct Projection
{
float rawProjection[6];
u32 type; // only GX_PERSPECTIVE or GX_ORTHOGRAPHIC are allowed
};
struct XFMemory
{
float posMatrices[256]; // 0x0000 - 0x00ff
u32 unk0[768]; // 0x0100 - 0x03ff
float normalMatrices[96]; // 0x0400 - 0x045f
u32 unk1[160]; // 0x0460 - 0x04ff
float postMatrices[256]; // 0x0500 - 0x05ff
Light lights[8]; // 0x0600 - 0x067f
u32 unk2[2432]; // 0x0680 - 0x0fff
u32 error; // 0x1000
u32 diag; // 0x1001
u32 state0; // 0x1002
u32 state1; // 0x1003
u32 xfClock; // 0x1004
u32 clipDisable; // 0x1005
u32 perf0; // 0x1006
u32 perf1; // 0x1007
INVTXSPEC hostinfo; // 0x1008 number of textures,colors,normals from vertex input
NumColorChannel numChan; // 0x1009
u32 ambColor[2]; // 0x100a, 0x100b
u32 matColor[2]; // 0x100c, 0x100d
LitChannel color[2]; // 0x100e, 0x100f
LitChannel alpha[2]; // 0x1010, 0x1011
DualTexInfo dualTexTrans; // 0x1012
u32 unk3; // 0x1013
u32 unk4; // 0x1014
u32 unk5; // 0x1015
u32 unk6; // 0x1016
u32 unk7; // 0x1017
TMatrixIndexA MatrixIndexA; // 0x1018
TMatrixIndexB MatrixIndexB; // 0x1019
Viewport viewport; // 0x101a - 0x101f
Projection projection; // 0x1020 - 0x1026
u32 unk8[24]; // 0x1027 - 0x103e
NumTexGen numTexGen; // 0x103f
TexMtxInfo texMtxInfo[8]; // 0x1040 - 0x1047
u32 unk9[8]; // 0x1048 - 0x104f
PostMtxInfo postMtxInfo[8]; // 0x1050 - 0x1057
};
extern XFMemory xfmem;
2014-11-27 15:53:11 -07:00
void LoadXFReg(u32 transferSize, u32 address, DataReader src);
void LoadIndexedXF(u32 val, int array);
Add the 'desynced GPU thread' mode. It's a relatively big commit (less big with -w), but it's hard to test any of this separately... The basic problem is that in netplay or movies, the state of the CPU must be deterministic, including when the game receives notification that the GPU has processed FIFO data. Dual core mode notifies the game whenever the GPU thread actually gets around to doing the work, so it isn't deterministic. Single core mode is because it notifies the game 'instantly' (after processing the data synchronously), but it's too slow for many systems and games. My old dc-netplay branch worked as follows: everything worked as normal except the state of the CP registers was a lie, and the CPU thread only delivered results when idle detection triggered (waiting for the GPU if they weren't ready at that point). Usually, a game is idle iff all the work for the frame has been done, except for a small amount of work depending on the GPU result, so neither the CPU or the GPU waiting on the other affected performance much. However, it's possible that the game could be waiting for some earlier interrupt, and any of several games which, for whatever reason, never went into a detectable idle (even when I tried to improve the detection) would never receive results at all. (The current method should have better compatibility, but it also has slightly higher overhead and breaks some other things, so I want to reimplement this, hopefully with less impact on the code, in the future.) With this commit, the basic idea is that the CPU thread acts as if the work has been done instantly, like single core mode, but actually hands it off asynchronously to the GPU thread (after backing up some data that the game might change in memory before it's actually done). Since the work isn't done, any feedback from the GPU to the CPU, such as real XFB/EFB copies (virtual are OK), EFB pokes, performance queries, etc. is broken; but most games work with these options disabled, and there is no need to try to detect what the CPU thread is doing. Technically: when the flag g_use_deterministic_gpu_thread (currently stuck on) is on, the CPU thread calls RunGpu like in single core mode. This function synchronously copies the data from the FIFO to the internal video buffer and updates the CP registers, interrupts, etc. However, instead of the regular ReadDataFromFifo followed by running the opcode decoder, it runs ReadDataFromFifoOnCPU -> OpcodeDecoder_Preprocess, which relatively quickly scans through the FIFO data, detects SetFinish calls etc., which are immediately fired, and saves certain associated data from memory (e.g. display lists) in AuxBuffers (a parallel stream to the main FIFO, which is a bit slow at the moment), before handing the data off to the GPU thread to actually render. That makes up the bulk of this commit. In various circumstances, including the aforementioned EFB pokes and performance queries as well as swap requests (i.e. the end of a frame - we don't want the CPU potentially pumping out frames too quickly and the GPU falling behind*), SyncGPU is called to wait for actual completion. The overhead mainly comes from OpcodeDecoder_Preprocess (which is, again, synchronous), as well as the actual copying. Currently, display lists and such are escrowed from main memory even though they usually won't change over the course of a frame, and textures are not even though they might, resulting in a small chance of graphical glitches. When the texture locking (i.e. fault on write) code lands, I can make this all correct and maybe a little faster. * This suggests an alternate determinism method of just delaying results until a short time before the end of each frame. For all I know this might mostly work - I haven't tried it - but if any significant work hinges on the competion of render to texture etc., the frame will be missed.
2014-08-27 20:56:19 -06:00
void PreprocessIndexedXF(u32 val, int refarray);