[D3D8 Patch] vertex shader integration + fixes - VS3
Raphaël Junqueira
fenix at club-internet.fr
Wed Dec 18 18:38:47 CST 2002
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi again,
this patch enable pixel shader rendering into D3D8 (and correct some stupid bugs i have)
Changelog:
- intergration with DrawPrimitive
- remove non-esthetic c++ comments into c comments block ;)
- more more debug
Problems:
- very slow
- i have problem with textures, if people can look at code ...
TODO:
- better integrate vertex shader with draw primitive (maybe we have to splt DrawPrimitive)
- optimize software emulation
- hardware vertex shader program support (using auto-detect caps)
- syntax validity of vertex shader declaration and vexter shader function code (Validate*Shader functions)
- vertex shader compilator (in D3DX8)
- pixel shaders
can people test this, using simple shaders programs please ?
thx, Raphael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)
iD8DBQE+ARUYp7NA3AmQTU4RAn+QAJ0V2R3CugfcYJlygJUMAOhJXj68bQCfQxfH
txo29ZkJSbuudmaxhYGUJSw=
=Hw4m
-----END PGP SIGNATURE-----
-------------- next part --------------
? tmp.diff
Index: d3d8_private.h
===================================================================
RCS file: /home/wine/wine/dlls/d3d8/d3d8_private.h,v
retrieving revision 1.8
diff -u -r1.8 d3d8_private.h
--- d3d8_private.h 18 Dec 2002 05:05:41 -0000 1.8
+++ d3d8_private.h 19 Dec 2002 00:28:25 -0000
@@ -198,6 +198,25 @@
SHADER8Data* data;
} PIXELSHADER8;
+/** temporary here waiting for buffer code */
+typedef struct vshader_input_data {
+ /*SHADER8Vector V[16];//0-15*/
+ SHADER8Vector V[16];
+} vshader_input_data;
+
+/** temporary here waiting for buffer code */
+typedef struct vshader_output_data {
+ SHADER8Vector oPos;
+ /*SHADER8Vector oD[2];//0-1*/
+ SHADER8Vector oD[2];
+ /*SHADER8Vector oT[4];//0-3*/
+ SHADER8Vector oT[4];
+ /*SHADER8Scalar oFog;*/
+ /*SHADER8Scalar oPts;*/
+ SHADER8Vector oFog;
+ SHADER8Vector oPts;
+} vshader_output_data;
+
/*
* External prototypes
*/
@@ -912,6 +931,6 @@
*/
DWORD vshader_decl_parse(VERTEXSHADER8* vshader);
DWORD vshader_program_parse(VERTEXSHADER8* vshader);
-
+BOOL vshader_program_execute_SW(VERTEXSHADER8* vshader, vshader_input_data* input, vshader_output_data* output);
#endif /* __WINE_D3DX8_PRIVATE_H */
Index: device.c
===================================================================
RCS file: /home/wine/wine/dlls/d3d8/device.c,v
retrieving revision 1.16
diff -u -r1.16 device.c
--- device.c 18 Dec 2002 05:05:41 -0000 1.16
+++ device.c 19 Dec 2002 00:28:40 -0000
@@ -53,6 +53,11 @@
memcpy(gl_mat, (mat), 16 * sizeof(float)); \
};
+#define VERTEX_SHADER(Handle) \
+ ((Handle <= VS_HIGHESTFIXEDFXF) ? ((Handle >= sizeof(VertexShaders) / sizeof(VERTEXSHADER8*)) ? NULL : VertexShaders[Handle]) : VertexShaders[Handle - VS_HIGHESTFIXEDFXF])
+
+#define TRACE_VECTOR(name) TRACE( #name "=(%f, %f, %f, %f)\n", name.x, name.y, name.z, name.w);
+
/*
* Globals
*/
@@ -68,6 +73,7 @@
0.0, 0.0, 0.0, 1.0
};
+
/* Routine common to the draw primitive and draw indexed primitive routines
Doesnt use gl pointer arrays as I dont believe we can support the blending
coordinates that way. */
@@ -89,17 +95,38 @@
int vx_index;
int NumVertexes = NumPrimitives;
+ VERTEXSHADER8* vertex_shader = NULL;
+ vshader_input_data vertex_shader_input;
ICOM_THIS(IDirect3DDevice8Impl,iface);
/* Dont understand how to handle multiple streams, but if a fixed
FVF is passed in rather than a handle, it must use stream 0 */
-
+
+ /*
if (This->StateBlock.VertexShader > VS_HIGHESTFIXEDFXF) {
FIXME("Cant handle created shaders yet\n");
return;
- } else {
+ } else
+ */
+
+ if (This->StateBlock.VertexShader > VS_HIGHESTFIXEDFXF) {
+ vertex_shader = VERTEX_SHADER(This->StateBlock.VertexShader);
+ if (NULL == vertex_shader) {
+ ERR("trying to use unitialised vertex shader: %lu\n", This->StateBlock.VertexShader);
+ return;
+ }
+ if (NULL == vertex_shader->function) {
+ TRACE("vertex shader declared without program, using FVF pure mode\n");
+ }
+ fvf = (D3DFORMAT) vertex_shader->fvf;
+ TRACE("vertex shader declared FVF: %x\n", vertex_shader->fvf);
+ memset(&vertex_shader_input, 0, sizeof(vshader_input_data));
+ }
+
+
+ {
int skip = This->StateBlock.stream_stride[0];
BOOL normal;
@@ -136,7 +163,7 @@
isPtSize = fvf & D3DFVF_PSIZE;
isDiffuse = fvf & D3DFVF_DIFFUSE;
isSpecular = fvf & D3DFVF_SPECULAR;
- numTextures = (fvf & D3DFVF_TEXCOUNT_MASK) >> D3DFVF_TEXCOUNT_SHIFT;
+ numTextures = (fvf & D3DFVF_TEXCOUNT_MASK) >> D3DFVF_TEXCOUNT_SHIFT;
TRACE("Drawing with FVF = %x, (n?%d, rhw?%d, ptSize(%d), diffuse?%d, specular?%d, numTextures=%d)\n",
fvf, normal, isRHW, isPtSize, isDiffuse, isSpecular, numTextures);
@@ -155,10 +182,10 @@
double height, width, minZ, maxZ;
/*
- * Already transformed vertex do not need transform
- * matrices. Reset all matrices to identity.
- * Leave the default matrix in world mode.
- */
+ * Already transformed vertex do not need transform
+ * matrices. Reset all matrices to identity.
+ * Leave the default matrix in world mode.
+ */
glMatrixMode(GL_PROJECTION);
checkGLcall("glMatrixMode");
glLoadIdentity();
@@ -259,11 +286,22 @@
curPos = curPos + sizeof(float);
TRACE("x,y,z=%f,%f,%f\n", x,y,z);
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vertex_shader_input.V[D3DVSDE_POSITION].x = x;
+ vertex_shader_input.V[D3DVSDE_POSITION].y = y;
+ vertex_shader_input.V[D3DVSDE_POSITION].z = z;
+ vertex_shader_input.V[D3DVSDE_POSITION].w = 1.0f;
+ }
+
/* RHW follows, only if transformed */
if (isRHW) {
rhw = *(float *)curPos;
curPos = curPos + sizeof(float);
TRACE("rhw=%f\n", rhw);
+
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vertex_shader_input.V[D3DVSDE_POSITION].w = rhw;
+ }
}
@@ -278,71 +316,173 @@
nz = *(float *)curPos;
curPos = curPos + sizeof(float);
TRACE("nx,ny,nz=%f,%f,%f\n", nx,ny,nz);
- }
+
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vertex_shader_input.V[D3DVSDE_NORMAL].x = nx;
+ vertex_shader_input.V[D3DVSDE_NORMAL].y = ny;
+ vertex_shader_input.V[D3DVSDE_NORMAL].z = nz;
+ vertex_shader_input.V[D3DVSDE_NORMAL].w = 1.0f;
+ }
+ }
if (isPtSize) {
ptSize = *(float *)curPos;
curPos = curPos + sizeof(float);
TRACE("ptSize=%f\n", ptSize);
+
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vertex_shader_input.V[D3DVSDE_PSIZE].x = ptSize;
+ vertex_shader_input.V[D3DVSDE_PSIZE].y = 0.0f;
+ vertex_shader_input.V[D3DVSDE_PSIZE].z = 0.0f;
+ vertex_shader_input.V[D3DVSDE_PSIZE].w = 1.0f;
+ }
}
if (isDiffuse) {
diffuseColor = *(DWORD *)curPos;
TRACE("diffuseColor=%lx\n", diffuseColor);
curPos = curPos + sizeof(DWORD);
+
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vertex_shader_input.V[D3DVSDE_DIFFUSE].x = (float) (((diffuseColor >> 16) & 0xFF) / 255.0f);
+ vertex_shader_input.V[D3DVSDE_DIFFUSE].y = (float) (((diffuseColor >> 8) & 0xFF) / 255.0f);
+ vertex_shader_input.V[D3DVSDE_DIFFUSE].z = (float) (((diffuseColor >> 0) & 0xFF) / 255.0f);
+ vertex_shader_input.V[D3DVSDE_DIFFUSE].w = (float) (((diffuseColor >> 24) & 0xFF) / 255.0f);
+ }
}
if (isSpecular) {
specularColor = *(DWORD *)curPos;
TRACE("specularColor=%lx\n", specularColor);
curPos = curPos + sizeof(DWORD);
+
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vertex_shader_input.V[D3DVSDE_SPECULAR].x = (float) (((specularColor >> 16) & 0xFF) / 255.0f);
+ vertex_shader_input.V[D3DVSDE_SPECULAR].y = (float) (((specularColor >> 8) & 0xFF) / 255.0f);
+ vertex_shader_input.V[D3DVSDE_SPECULAR].z = (float) (((specularColor >> 0) & 0xFF) / 255.0f);
+ vertex_shader_input.V[D3DVSDE_SPECULAR].w = (float) (((specularColor >> 24) & 0xFF) / 255.0f);
+ }
}
/* ToDo: Texture coords */
- for (textureNo = 0;textureNo<numTextures; textureNo++) {
+ for (textureNo = 0; textureNo < numTextures; ++textureNo) {
float s,t,r,q;
/* Query tex coords */
if (This->StateBlock.textures[textureNo] != NULL) {
- switch (IDirect3DBaseTexture8Impl_GetType((LPDIRECT3DBASETEXTURE8) This->StateBlock.textures[textureNo])) {
- case D3DRTYPE_TEXTURE:
- s = *(float *)curPos;
- curPos = curPos + sizeof(float);
- t = *(float *)curPos;
- curPos = curPos + sizeof(float);
- TRACE("tex:%d, s,t=%f,%f\n", textureNo, s,t);
- glMultiTexCoord2fARB(GL_TEXTURE0_ARB + textureNo, s, t);
- break;
-
- case D3DRTYPE_VOLUMETEXTURE:
- s = *(float *)curPos;
- curPos = curPos + sizeof(float);
- t = *(float *)curPos;
- curPos = curPos + sizeof(float);
- r = *(float *)curPos;
- curPos = curPos + sizeof(float);
- TRACE("tex:%d, s,t,r=%f,%f,%f\n", textureNo, s,t,r);
- glMultiTexCoord3fARB(GL_TEXTURE0_ARB + textureNo, s, t, r);
- break;
-
- default:
- r=0;q=0; /* Avoid compiler warnings, need these vars later for other textures */
- FIXME("Unhandled texture type\n");
- }
+ switch (IDirect3DBaseTexture8Impl_GetType((LPDIRECT3DBASETEXTURE8) This->StateBlock.textures[textureNo])) {
+ case D3DRTYPE_TEXTURE:
+ s = *(float *)curPos;
+ curPos = curPos + sizeof(float);
+ t = *(float *)curPos;
+ curPos = curPos + sizeof(float);
+ TRACE("tex:%d, s,t=%f,%f\n", textureNo, s,t);
+
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vertex_shader_input.V[D3DVSDE_TEXCOORD0 + textureNo].x = s;
+ vertex_shader_input.V[D3DVSDE_TEXCOORD0 + textureNo].y = t;
+ vertex_shader_input.V[D3DVSDE_TEXCOORD0 + textureNo].z = 0.0f;
+ vertex_shader_input.V[D3DVSDE_TEXCOORD0 + textureNo].w = 1.0f;
+ } else {
+ glMultiTexCoord2fARB(GL_TEXTURE0_ARB + textureNo, s, t);
+ }
+ break;
+
+ case D3DRTYPE_VOLUMETEXTURE:
+ s = *(float *)curPos;
+ curPos = curPos + sizeof(float);
+ t = *(float *)curPos;
+ curPos = curPos + sizeof(float);
+ r = *(float *)curPos;
+ curPos = curPos + sizeof(float);
+ TRACE("tex:%d, s,t,r=%f,%f,%f\n", textureNo, s,t,r);
+
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vertex_shader_input.V[D3DVSDE_TEXCOORD0 + textureNo].x = s;
+ vertex_shader_input.V[D3DVSDE_TEXCOORD0 + textureNo].y = t;
+ vertex_shader_input.V[D3DVSDE_TEXCOORD0 + textureNo].z = r;
+ vertex_shader_input.V[D3DVSDE_TEXCOORD0 + textureNo].w = 1.0f;
+ } else {
+ glMultiTexCoord3fARB(GL_TEXTURE0_ARB + textureNo, s, t, r);
+ }
+ break;
+
+ default:
+ r = 0.0f; q = 0.0f; /* Avoid compiler warnings, need these vars later for other textures */
+ FIXME("Unhandled texture type\n");
+ }
} else {
- /* Note I have seen a program actually do this, so just hide it and continue */
- TRACE("Very odd - texture requested in FVF but not bound!\n");
+ /* Note I have seen a program actually do this, so just hide it and continue */
+ TRACE("Very odd - texture requested in FVF but not bound!\n");
}
}
+ /** if vertex shader specified ... using it */
+ if (NULL != vertex_shader && NULL != vertex_shader->function) {
+ vshader_output_data vs_o;
+ memset(&vs_o, 0, sizeof(vshader_output_data));
+ vshader_program_execute_SW(vertex_shader, &vertex_shader_input, &vs_o);
+
+ x = vs_o.oPos.x;
+ y = vs_o.oPos.y;
+ z = vs_o.oPos.z;
+
+ if (isRHW) {
+ rhw = vs_o.oPos.w;
+ }
+ TRACE_VECTOR(vs_o.oPos);
+ if (isDiffuse) {
+ diffuseColor = D3DCOLOR_COLORVALUE(vs_o.oD[0].x, vs_o.oD[0].y, vs_o.oD[0].z, vs_o.oD[0].w);
+ TRACE_VECTOR(vs_o.oD[0]);
+ }
+ if (isSpecular) {
+ specularColor = D3DCOLOR_COLORVALUE(vs_o.oD[1].x, vs_o.oD[1].y, vs_o.oD[1].z, vs_o.oD[1].w);
+ TRACE_VECTOR(vs_o.oD[1]);
+ }
+
+ /** TODO: reupdate textures coords binding using vs_o.oT[0->3] */
+ for (textureNo = 0; textureNo < min(numTextures, 4); ++textureNo) {
+ float s, t, r, q;
+ /* Query tex coords */
+ if (This->StateBlock.textures[textureNo] != NULL) {
+ switch (IDirect3DBaseTexture8Impl_GetType((LPDIRECT3DBASETEXTURE8) This->StateBlock.textures[textureNo])) {
+ case D3DRTYPE_TEXTURE:
+ TRACE_VECTOR(vs_o.oT[textureNo]);
+ s = vs_o.oT[textureNo].x;
+ t = vs_o.oT[textureNo].y;
+ TRACE("tex:%d, s,t=%f,%f\n", textureNo, s,t);
+ glMultiTexCoord2fARB(GL_TEXTURE0_ARB + textureNo, s, t);
+ break;
+
+ case D3DRTYPE_VOLUMETEXTURE:
+ TRACE_VECTOR(vs_o.oT[textureNo]);
+ s = vs_o.oT[textureNo].x;
+ t = vs_o.oT[textureNo].y;
+ r = vs_o.oT[textureNo].z;
+ TRACE("tex:%d, s,t,r=%f,%f,%f\n", textureNo, s,t,r);
+ glMultiTexCoord3fARB(GL_TEXTURE0_ARB + textureNo, s, t, r);
+ break;
+
+ default:
+ /* Avoid compiler warnings, need these vars later for other textures */
+ r = 0.0f; q = 0.0f;
+ FIXME("Unhandled texture type\n");
+ }
+ }
+ }
+ }
+
/* Handle these vertexes */
if (isDiffuse) {
glColor4f(((diffuseColor >> 16) & 0xFF) / 255.0,
((diffuseColor >> 8) & 0xFF) / 255.0,
((diffuseColor >> 0) & 0xFF) / 255.0,
((diffuseColor >> 24) & 0xFF) / 255.0);
- TRACE("glColor4f: r,g,b,a=%f,%f,%f,%f\n", ((diffuseColor >> 16) & 0xFF) / 255.0, ((diffuseColor >> 8) & 0xFF) / 255.0,
- ((diffuseColor >> 0) & 0xFF) / 255.0, ((diffuseColor >> 24) & 0xFF) / 255.0);
+ TRACE("glColor4f: r,g,b,a=%f,%f,%f,%f\n",
+ ((diffuseColor >> 16) & 0xFF) / 255.0,
+ ((diffuseColor >> 8) & 0xFF) / 255.0,
+ ((diffuseColor >> 0) & 0xFF) / 255.0,
+ ((diffuseColor >> 24) & 0xFF) / 255.0);
}
if (normal) {
@@ -3152,19 +3292,14 @@
object->decl = pDeclaration;
object->function = pFunction;
- /*
- for (i = 0; 0xFFFFFFFF != pDeclaration[i]; ++i) ;
- object->declLength = i + 1;
- if (NULL != pFunction) {
- for (i = 0; 0xFFFFFFFF != pFunction[i]; ++i) ;
- object->functionLength = i + 1;
- } else {
- object->functionLength = 1; // no Function defined use fixed function vertex processing
- }
- */
vshader_decl_parse(object);
vshader_program_parse(object);
+ /* copy the function ... because it will certainly be released by application */
+ if (NULL != pFunction) {
+ object->function = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, object->functionLength);
+ memcpy(object->function, pFunction, object->functionLength);
+ }
return D3D_OK;
}
HRESULT WINAPI IDirect3DDevice8Impl_SetVertexShader(LPDIRECT3DDEVICE8 iface, DWORD Handle) {
@@ -3212,8 +3347,6 @@
VertexShaders[Handle - VS_HIGHESTFIXEDFXF] = NULL;
return D3D_OK;
}
-
-#define VERTEX_SHADER(Handle) ((Handle <= VS_HIGHESTFIXEDFXF) ? ((Handle >= sizeof(VertexShaders) / sizeof(VERTEXSHADER8*)) ? NULL : VertexShaders[Handle]) : VertexShaders[Handle - VS_HIGHESTFIXEDFXF])
HRESULT WINAPI IDirect3DDevice8Impl_SetVertexShaderConstant(LPDIRECT3DDEVICE8 iface, DWORD Register, CONST void* pConstantData, DWORD ConstantCount) {
ICOM_THIS(IDirect3DDevice8Impl,iface);
Index: directx.c
===================================================================
RCS file: /home/wine/wine/dlls/d3d8/directx.c,v
retrieving revision 1.8
diff -u -r1.8 directx.c
--- directx.c 18 Dec 2002 05:05:41 -0000 1.8
+++ directx.c 19 Dec 2002 00:28:42 -0000
@@ -584,11 +584,13 @@
LEAVE_GL();
- {
+
+ { /* Print some caps infos */
GLint gl_max_texture_units_arb;
glGetIntegerv(GL_MAX_TEXTURE_UNITS_ARB, &gl_max_texture_units_arb);
TRACE("GLCaps: GL_MAX_TEXTURE_UNITS_ARB=%d\n", gl_max_texture_units_arb);
}
+
{ /* Set a default viewport */
D3DVIEWPORT8 vp;
Index: shader.c
===================================================================
RCS file: /home/wine/wine/dlls/d3d8/shader.c,v
retrieving revision 1.2
diff -u -r1.2 shader.c
--- shader.c 18 Dec 2002 05:05:41 -0000 1.2
+++ shader.c 19 Dec 2002 00:28:45 -0000
@@ -63,22 +63,39 @@
*/
} shader_opcode;
-typedef struct vshader_input_data {
- /*SHADER8Vector V[16];//0-15*/
- SHADER8Vector V[16];
-} vshader_input_data;
-
-typedef struct vshader_output_data {
- SHADER8Vector oPos;
- /*SHADER8Vector oD[2];//0-1*/
- SHADER8Vector oD[2];
- /*SHADER8Vector oT[4];//0-3*/
- SHADER8Vector oT[4];
- /*SHADER8Scalar oFog;*/
- /*SHADER8Scalar oPts;*/
- SHADER8Vector oFog;
- SHADER8Vector oPts;
-} vshader_output_data;
+/** Vertex Shader Declaration data types tokens */
+static CONST char* VertexShaderDeclDataTypes [] = {
+ "D3DVSDT_FLOAT1",
+ "D3DVSDT_FLOAT2",
+ "D3DVSDT_FLOAT3",
+ "D3DVSDT_FLOAT4",
+ "D3DVSDT_D3DCOLOR",
+ "D3DVSDT_UBYTE4",
+ "D3DVSDT_SHORT2",
+ "D3DVSDT_SHORT4",
+ NULL
+};
+
+static CONST char* VertexShaderDeclRegister [] = {
+ "D3DVSDE_POSITION",
+ "D3DVSDE_BLENDWEIGHT",
+ "D3DVSDE_BLENDINDICES",
+ "D3DVSDE_NORMAL",
+ "D3DVSDE_PSIZE",
+ "D3DVSDE_DIFFUSE",
+ "D3DVSDE_SPECULAR",
+ "D3DVSDE_TEXCOORD0",
+ "D3DVSDE_TEXCOORD1",
+ "D3DVSDE_TEXCOORD2",
+ "D3DVSDE_TEXCOORD3",
+ "D3DVSDE_TEXCOORD4",
+ "D3DVSDE_TEXCOORD5",
+ "D3DVSDE_TEXCOORD6",
+ "D3DVSDE_TEXCOORD7",
+ "D3DVSDE_POSITION2",
+ "D3DVSDE_NORMAL2",
+ NULL
+};
/*******************************
* vshader functions software VM
@@ -97,6 +114,11 @@
void vshader_dp4(SHADER8Vector* d, SHADER8Vector* s0, SHADER8Vector* s1) {
d->x = d->y = d->z = d->w = s0->x * s1->x + s0->y * s1->y + s0->z * s1->z + s0->w * s1->w;
+
+ /*
+ DPRINTF("executing dp4: s0=(%f, %f, %f, %f) s1=(%f, %f, %f, %f) => d=(%f, %f, %f, %f)\n",
+ s0->x, s0->y, s0->z, s0->w, s1->x, s1->y, s1->z, s1->w, d->x, d->y, d->z, d->w);
+ */
}
void vshader_dst(SHADER8Vector* d, SHADER8Vector* s0, SHADER8Vector* s1) {
@@ -104,6 +126,11 @@
d->y = s0->y * s1->y;
d->z = s0->z;
d->w = s1->w;
+
+ /*
+ DPRINTF("executing dst: s0=(%f, %f, %f, %f) s1=(%f, %f, %f, %f) => d=(%f, %f, %f, %f)\n",
+ s0->x, s0->y, s0->z, s0->w, s1->x, s1->y, s1->z, s1->w, d->x, d->y, d->z, d->w);
+ */
}
void vshader_expp(SHADER8Vector* d, SHADER8Vector* s0) {
@@ -112,6 +139,11 @@
d->y = s0->w - tmp_f;
d->z = pow(2, s0->w);
d->w = 1;
+
+ /*
+ DPRINTF("executing exp: s0=(%f, %f, %f, %f) => d=(%f, %f, %f, %f)\n",
+ s0->x, s0->y, s0->z, s0->w, d->x, d->y, d->z, d->w);
+ */
}
void vshader_lit(SHADER8Vector* d, SHADER8Vector* s0) {
@@ -119,6 +151,11 @@
d->y = (0 < s0->x) ? s0->x : 0;
d->z = (0 < s0->x && 0 < s0->y) ? pow(s0->y, s0->w) : 0;
d->w = 1;
+
+ /*
+ DPRINTF("executing lit: s0=(%f, %f, %f, %f) => d=(%f, %f, %f, %f)\n",
+ s0->x, s0->y, s0->z, s0->w, d->x, d->y, d->z, d->w);
+ */
}
void vshader_logp(SHADER8Vector* d, SHADER8Vector* s0) {
@@ -151,6 +188,11 @@
d->y = s0->y;
d->z = s0->z;
d->w = s0->w;
+
+ /*
+ DPRINTF("executing mov: s0=(%f, %f, %f, %f) => d=(%f, %f, %f, %f)\n",
+ s0->x, s0->y, s0->z, s0->w, d->x, d->y, d->z, d->w);
+ */
}
void vshader_mul(SHADER8Vector* d, SHADER8Vector* s0, SHADER8Vector* s1) {
@@ -158,6 +200,11 @@
d->y = s0->y * s1->y;
d->z = s0->z * s1->z;
d->w = s0->w * s1->w;
+
+ /*
+ DPRINTF("executing mul: s0=(%f, %f, %f, %f) s1=(%f, %f, %f, %f) => d=(%f, %f, %f, %f)\n",
+ s0->x, s0->y, s0->z, s0->w, s1->x, s1->y, s1->z, s1->w, d->x, d->y, d->z, d->w);
+ */
}
void vshader_nop(void) {
@@ -223,6 +270,7 @@
const shader_opcode* vshader_program_get_opcode(const DWORD code) {
DWORD i = 0;
+ /** TODO: use dichotomic search */
while (NULL != vshader_ins[i].name) {
if ((code & D3DSI_OPCODE_MASK) == vshader_ins[i].opcode) {
return &vshader_ins[i];
@@ -339,7 +387,7 @@
DPRINTF("\n");
}
}
- vshader->functionLength = len * sizeof(DWORD);
+ vshader->functionLength = (len + 1) * sizeof(DWORD);
} else {
vshader->functionLength = 1; /* no Function defined use fixed function vertex processing */
}
@@ -347,7 +395,7 @@
}
BOOL vshader_program_execute_HAL(VERTEXSHADER8* vshader,
- const vshader_input_data* input,
+ vshader_input_data* input,
vshader_output_data* output) {
/**
* TODO: use the NV_vertex_program (or 1_1) extension
@@ -356,6 +404,8 @@
return TRUE;
}
+#define TRACE_VECTOR(name) DPRINTF( #name "=(%f, %f, %f, %f)\n", name.x, name.y, name.z, name.w)
+
BOOL vshader_program_execute_SW(VERTEXSHADER8* vshader,
vshader_input_data* input,
vshader_output_data* output) {
@@ -372,35 +422,52 @@
/** functions parameters */
SHADER8Vector* p[4];
SHADER8Vector* p_send[4];
-
DWORD i;
+ /* vshader_program_parse(vshader); */
+ /*
+ TRACE_VECTOR(vshader->data->C[0]);
+ TRACE_VECTOR(vshader->data->C[1]);
+ TRACE_VECTOR(vshader->data->C[2]);
+ TRACE_VECTOR(vshader->data->C[3]);
+ TRACE_VECTOR(input->V[D3DVSDE_POSITION]);
+ TRACE_VECTOR(input->V[D3DVSDE_BLENDWEIGHT]);
+ TRACE_VECTOR(input->V[D3DVSDE_BLENDINDICES]);
+ TRACE_VECTOR(input->V[D3DVSDE_NORMAL]);
+ TRACE_VECTOR(input->V[D3DVSDE_PSIZE]);
+ TRACE_VECTOR(input->V[D3DVSDE_DIFFUSE]);
+ TRACE_VECTOR(input->V[D3DVSDE_SPECULAR]);
+ TRACE_VECTOR(input->V[D3DVSDE_TEXCOORD0]);
+ TRACE_VECTOR(input->V[D3DVSDE_TEXCOORD1]);
+ */
+
/* the first dword is the version tag */
/* TODO: parse it */
- ++pToken;
while (D3DVS_END() != *pToken) {
curOpcode = vshader_program_get_opcode(*pToken);
++pToken;
if (NULL == curOpcode) {
/* unkown current opcode ... */
while (*pToken & 0x80000000) {
- DPRINTF("unrecognized opcode: %08lX\n", *pToken);
+ DPRINTF("unrecognized opcode: pos=%d token=%08lX\n", pToken - vshader->function, *pToken);
++pToken;
}
/*return FALSE;*/
} else {
- if (curOpcode->num_params > 0) {
-
+ if (curOpcode->num_params > 0) {
+ /*DPRINTF(">> execting opcode: pos=%lu opcode_name=%s token=%08lX\n", pToken - vshader->function, curOpcode->name, *pToken);*/
for (i = 0; i < curOpcode->num_params; ++i) {
DWORD reg = pToken[i] & 0x00001FFF;
DWORD regtype = ((pToken[i] & D3DSP_REGTYPE_MASK) >> D3DSP_REGTYPE_SHIFT);
switch (regtype << D3DSP_REGTYPE_SHIFT) {
case D3DSPR_TEMP:
+ /*DPRINTF("p[%d]=R[%d]\n", i, reg);*/
p[i] = &R[reg];
break;
case D3DSPR_INPUT:
+ /*DPRINTF("p[%d]=V[%s]\n", i, VertexShaderDeclRegister[reg]);*/
p[i] = &input->V[reg];
break;
case D3DSPR_CONST:
@@ -411,8 +478,11 @@
}
break;
case D3DSPR_ADDR: /*case D3DSPR_TEXTURE:*/
- if (0 != reg)
- ERR("cannot handle address registers != a0");
+ if (0 != reg) {
+ ERR("cannot handle address registers != a0, forcing use of a0\n");
+ reg = 0;
+ }
+ /*DPRINTF("p[%d]=A[%d]\n", i, reg);*/
p[i] = &A[reg];
break;
case D3DSPR_RASTOUT:
@@ -429,30 +499,45 @@
}
break;
case D3DSPR_ATTROUT:
+ /*DPRINTF("p[%d]=oD[%d]\n", i, reg);*/
p[i] = &output->oD[reg];
break;
case D3DSPR_TEXCRDOUT:
+ /*DPRINTF("p[%d]=oT[%d]\n", i, reg);*/
p[i] = &output->oT[reg];
break;
default:
break;
}
- if (i > 1) { /* input reg */
+ if (i > 0) { /* input reg */
DWORD swizzle = (pToken[i] & D3DVS_SWIZZLE_MASK) >> D3DVS_SWIZZLE_SHIFT;
DWORD swizzle_x = swizzle & 0x03;
DWORD swizzle_y = (swizzle >> 2) & 0x03;
DWORD swizzle_z = (swizzle >> 4) & 0x03;
DWORD swizzle_w = (swizzle >> 6) & 0x03;
if ((D3DVS_NOSWIZZLE >> D3DVS_SWIZZLE_SHIFT) == swizzle) {
+ /*DPRINTF("p[%d] not swizzled\n", i);*/
p_send[i] = p[i];
} else {
- float* tt = (float*) p[i];
- s[i].x = tt[swizzle_x];
- s[i].y = tt[swizzle_y];
- s[i].z = tt[swizzle_z];
- s[i].w = tt[swizzle_w];
- p_send[i] = &s[i];
+ /*DPRINTF("p[%d] swizzled\n", i);*/
+ if (swizzle_x == swizzle_y &&
+ swizzle_x == swizzle_z &&
+ swizzle_x == swizzle_w) {
+ float* tt = (float*) p[i];
+ s[i].x = tt[swizzle_x];
+ s[i].y = 0.0f;
+ s[i].z = 0.0f;
+ s[i].w = 0.0f;
+ p_send[i] = &s[i];
+ } else {
+ float* tt = (float*) p[i];
+ s[i].x = tt[swizzle_x];
+ s[i].y = tt[swizzle_y];
+ s[i].z = tt[swizzle_z];
+ s[i].w = tt[swizzle_w];
+ p_send[i] = &s[i];
+ }
}
} else { /* output reg */
if ((pToken[i] & D3DSP_WRITEMASK_ALL) == D3DSP_WRITEMASK_ALL) {
@@ -491,7 +576,20 @@
if (pToken[0] & D3DSP_WRITEMASK_2) p[0]->z = d.z;
if (pToken[0] & D3DSP_WRITEMASK_3) p[0]->w = d.w;
}
-
+
+ /*
+ TRACE_VECTOR(output->oPos);
+ TRACE_VECTOR(output->oD[0]);
+ TRACE_VECTOR(output->oD[1]);
+ TRACE_VECTOR(output->oT[0]);
+ TRACE_VECTOR(output->oT[1]);
+ TRACE_VECTOR(R[0]);
+ TRACE_VECTOR(R[1]);
+ TRACE_VECTOR(R[2]);
+ TRACE_VECTOR(R[3]);
+ TRACE_VECTOR(R[4]);
+ */
+
/* to next opcode token */
pToken += curOpcode->num_params;
}
@@ -503,41 +601,6 @@
* Vertex Shader Declaration Parser First draft ...
*/
-/** Vertex Shader Declaration data types tokens */
-static CONST char* VertexShaderDeclDataTypes [] = {
- "D3DVSDT_FLOAT1",
- "D3DVSDT_FLOAT2",
- "D3DVSDT_FLOAT3",
- "D3DVSDT_FLOAT4",
- "D3DVSDT_D3DCOLOR",
- "D3DVSDT_UBYTE4",
- "D3DVSDT_SHORT2",
- "D3DVSDT_SHORT4",
- NULL
-};
-
-static CONST char* VertexShaderDeclRegister [] = {
- "D3DVSDE_POSITION",
- "D3DVSDE_BLENDWEIGHT",
- "D3DVSDE_BLENDINDICES",
- "D3DVSDE_NORMAL",
- "D3DVSDE_PSIZE",
- "D3DVSDE_DIFFUSE",
- "D3DVSDE_SPECULAR",
- "D3DVSDE_TEXCOORD0",
- "D3DVSDE_TEXCOORD1",
- "D3DVSDE_TEXCOORD2",
- "D3DVSDE_TEXCOORD3",
- "D3DVSDE_TEXCOORD4",
- "D3DVSDE_TEXCOORD5",
- "D3DVSDE_TEXCOORD6",
- "D3DVSDE_TEXCOORD7",
- "D3DVSDE_POSITION2",
- "D3DVSDE_NORMAL2",
- NULL
-};
-
-
/** todo check decl validity */
DWORD vshader_decl_parse_token(const DWORD* pToken) {
const DWORD token = *pToken;
@@ -622,6 +685,7 @@
DWORD token;
DWORD tokenlen;
DWORD tokentype;
+ DWORD tex = D3DFVF_TEX0;
while (D3DVSD_END() != *pToken) {
token = *pToken;
@@ -632,8 +696,17 @@
if (D3DVSD_TOKEN_STREAMDATA == tokentype && 0 == (0x10000000 & tokentype)) {
DWORD type = ((token & D3DVSD_DATATYPEMASK) >> D3DVSD_DATATYPESHIFT);
DWORD reg = ((token & D3DVSD_VERTEXREGMASK) >> D3DVSD_VERTEXREGSHIFT);
+
switch (reg) {
- case D3DVSDE_POSITION: fvf |= D3DFVF_XYZ; break;
+ case D3DVSDE_POSITION:
+ switch (type) {
+ case D3DVSDT_FLOAT3: fvf |= D3DFVF_XYZ; break;
+ case D3DVSDT_FLOAT4: fvf |= D3DFVF_XYZRHW; break;
+ default:
+ /** errooooorr what to do ? */
+ ERR("Error in VertexShader declaration of D3DVSDE_POSITION register: unsupported type %lu\n", type);
+ }
+ break;
case D3DVSDE_BLENDWEIGHT:
switch (type) {
case D3DVSDT_FLOAT1: fvf |= D3DFVF_XYZB1; break;
@@ -646,31 +719,40 @@
}
break;
- case D3DVSDE_BLENDINDICES: fvf |= D3DFVF_LASTBETA_UBYTE4; break;
- case D3DVSDE_NORMAL: fvf |= D3DFVF_NORMAL; break;
- case D3DVSDE_PSIZE: fvf |= D3DFVF_PSIZE; break;
- case D3DVSDE_DIFFUSE: fvf |= D3DFVF_DIFFUSE; break;
- case D3DVSDE_SPECULAR: fvf |= D3DFVF_SPECULAR; break;
- case D3DVSDE_TEXCOORD0: fvf |= D3DFVF_TEX1; break;
- case D3DVSDE_TEXCOORD1: fvf |= D3DFVF_TEX2; break;
- case D3DVSDE_TEXCOORD2: fvf |= D3DFVF_TEX3; break;
- case D3DVSDE_TEXCOORD3: fvf |= D3DFVF_TEX4; break;
- case D3DVSDE_TEXCOORD4: fvf |= D3DFVF_TEX5; break;
- case D3DVSDE_TEXCOORD5: fvf |= D3DFVF_TEX6; break;
- case D3DVSDE_TEXCOORD6: fvf |= D3DFVF_TEX7; break;
- case D3DVSDE_TEXCOORD7: fvf |= D3DFVF_TEX8; break;
+ case D3DVSDE_BLENDINDICES: fvf |= D3DFVF_LASTBETA_UBYTE4; break; /* TODO: undertand ;( */
+ case D3DVSDE_NORMAL: fvf |= D3DFVF_NORMAL; break; /* TODO: only FLOAT3 supported ... another choice possible ? */
+ case D3DVSDE_PSIZE: fvf |= D3DFVF_PSIZE; break; /* TODO: only FLOAT1 supported ... another choice possible ? */
+ case D3DVSDE_DIFFUSE: fvf |= D3DFVF_DIFFUSE; break; /* TODO: only FLOAT1 supported */
+ case D3DVSDE_SPECULAR: fvf |= D3DFVF_SPECULAR; break; /* TODO: only FLOAT1 supported */
+ /**
+ * TODO: for TEX* only FLOAT2 supported
+ * by default using texture type info
+ */
+ case D3DVSDE_TEXCOORD0: tex = max(tex, D3DFVF_TEX1); break;
+ case D3DVSDE_TEXCOORD1: tex = max(tex, D3DFVF_TEX2); break;
+ case D3DVSDE_TEXCOORD2: tex = max(tex, D3DFVF_TEX3); break;
+ case D3DVSDE_TEXCOORD3: tex = max(tex, D3DFVF_TEX4); break;
+ case D3DVSDE_TEXCOORD4: tex = max(tex, D3DFVF_TEX5); break;
+ case D3DVSDE_TEXCOORD5: tex = max(tex, D3DFVF_TEX6); break;
+ case D3DVSDE_TEXCOORD6: tex = max(tex, D3DFVF_TEX7); break;
+ case D3DVSDE_TEXCOORD7: tex = max(tex, D3DFVF_TEX8); break;
case D3DVSDE_POSITION2: /* maybe D3DFVF_XYZRHW instead D3DFVF_XYZ (of D3DVDE_POSITION) ... to see */
case D3DVSDE_NORMAL2: /* FIXME i don't know what to do here ;( */
FIXME("[%lu] registers in VertexShader declaration not supported yet (token:0x%08lx)\n", reg, token);
break;
}
+ /*TRACE("VertexShader declaration define %x as current FVF\n", fvf);*/
}
len += tokenlen;
pToken += tokenlen;
}
+ if (tex > 0) {
+ /*TRACE("VertexShader declaration define %x as texture level\n", tex);*/
+ fvf |= tex;
+ }
/* here D3DVSD_END() */
len += vshader_decl_parse_token(pToken);
- if (NULL == vshader->function) vshader->fvf = fvf;
+ vshader->fvf = fvf;
vshader->declLength = len * sizeof(DWORD);
return len * sizeof(DWORD);
}
@@ -684,3 +766,4 @@
FIXME("(void): stub\n");
return 0;
}
+
More information about the wine-devel
mailing list