The-Homebrew-Cloud/dolphin
2
0
Fork 0
mirror of https://github.com/dolphin-emu/dolphin.git synced 2025-03-12 14:46:49 +01:00
Code Issues Packages Projects Releases Wiki Activity

Readability, comments, tiny typo bugfix in gl gfx plugin (fix is on line 322 in pixelshadermanager)

Browse Source 
git-svn-id: https://dolphin-emu.googlecode.com/svn/trunk@930 8ced0084-cf51-0410-be5f-012b33b47a6e
This commit is contained in:
hrydgard 2008-10-21 20:34:56 +00:00
parent e1baf2ead4
commit 58e3c8df8f
6 changed files with 218 additions and 185 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
Source/Plugins/Plugin_VideoOGL/Src/BPStructs.cpp
View File

@ -161,12 +161,12 @@ void BPWritten(int addr, int changes, int newval)
case BPMEM_LINEPTWIDTH: case BPMEM_LINEPTWIDTH:
{ {
float fratio = VertexShaderMngr::rawViewport[0] != 0 ? (float)Renderer::GetTargetWidth() / 640.0f : 1.0f; float fratio = VertexShaderMngr::GetPixelAspectRatio();
if (bpmem.lineptwidth.linesize > 0) if (bpmem.lineptwidth.linesize > 0)
glLineWidth((float)bpmem.lineptwidth.linesize * fratio / 6.0f); // scale by ratio of widths glLineWidth((float)bpmem.lineptwidth.linesize * fratio / 6.0f); // scale by ratio of widths
if (bpmem.lineptwidth.pointsize > 0) if (bpmem.lineptwidth.pointsize > 0)
glPointSize((float)bpmem.lineptwidth.pointsize * fratio / 6.0f); glPointSize((float)bpmem.lineptwidth.pointsize * fratio / 6.0f);
break; break;
} }
case 0x43: case 0x43:

105
Source/Plugins/Plugin_VideoOGL/Src/PixelShaderManager.cpp
View File

@ -133,8 +133,7 @@ FRAGMENTSHADER* PixelShaderMngr::GetShader()
} }
PSCacheEntry& newentry = pshaders[uid]; PSCacheEntry& newentry = pshaders[uid];
char *code = GeneratePixelShader(s_texturemask,
char *code = GeneratePixelShader(s_texturemask,
Renderer::GetZBufferTarget() != 0, Renderer::GetZBufferTarget() != 0,
Renderer::GetRenderMode() != Renderer::RM_Normal); Renderer::GetRenderMode() != Renderer::RM_Normal);
@ -184,12 +183,9 @@ void PixelShaderMngr::Cleanup()
bool PixelShaderMngr::CompilePixelShader(FRAGMENTSHADER& ps, const char* pstrprogram) bool PixelShaderMngr::CompilePixelShader(FRAGMENTSHADER& ps, const char* pstrprogram)
{ {
// todo: Should error be checked?
// CGerror cgerr = cgGetError();
char stropt[64]; char stropt[64];
sprintf(stropt, "MaxLocalParams=32,NumInstructionSlots=%d", s_nMaxPixelInstructions); sprintf(stropt, "MaxLocalParams=32,NumInstructionSlots=%d", s_nMaxPixelInstructions);
const char* opts[] = {"-profileopts",stropt,"-O2","-q",NULL}; const char* opts[] = {"-profileopts", stropt, "-O2", "-q", NULL};
CGprogram tempprog = cgCreateProgram(g_cgcontext, CG_SOURCE, pstrprogram, g_cgfProf, "main", opts); CGprogram tempprog = cgCreateProgram(g_cgcontext, CG_SOURCE, pstrprogram, g_cgfProf, "main", opts);
if (!cgIsProgram(tempprog) || cgGetError() != CG_NO_ERROR) { if (!cgIsProgram(tempprog) || cgGetError() != CG_NO_ERROR) {
ERROR_LOG("Failed to create ps %s:\n", cgGetLastListing(g_cgcontext)); ERROR_LOG("Failed to create ps %s:\n", cgGetLastListing(g_cgcontext));
@ -197,9 +193,9 @@ bool PixelShaderMngr::CompilePixelShader(FRAGMENTSHADER& ps, const char* pstrpro
return false; return false;
} }
char* pcompiledprog = (char*)cgGetProgramString(tempprog, CG_COMPILED_PROGRAM); char *pcompiledprog = (char*)cgGetProgramString(tempprog, CG_COMPILED_PROGRAM);
char* plocal = strstr(pcompiledprog, "program.local"); char *plocal = strstr(pcompiledprog, "program.local");
while ( plocal != NULL) { while (plocal != NULL) {
const char* penv = " program.env"; const char* penv = " program.env";
memcpy(plocal, penv, 13); memcpy(plocal, penv, 13);
plocal = strstr(plocal+13, "program.local"); plocal = strstr(plocal+13, "program.local");
@ -207,6 +203,7 @@ bool PixelShaderMngr::CompilePixelShader(FRAGMENTSHADER& ps, const char* pstrpro
if (Renderer::IsUsingATIDrawBuffers()) { if (Renderer::IsUsingATIDrawBuffers()) {
// sometimes compilation can use ARB_draw_buffers, which would fail for ATI cards // sometimes compilation can use ARB_draw_buffers, which would fail for ATI cards
// replace the three characters ARB with ATI. TODO - check whether this is fixed in modern ATI drivers.
char* poptions = strstr(pcompiledprog, "ARB_draw_buffers"); char* poptions = strstr(pcompiledprog, "ARB_draw_buffers");
if (poptions != NULL) { if (poptions != NULL) {
poptions[0] = 'A'; poptions[0] = 'A';
@ -313,6 +310,7 @@ void PixelShaderMngr::SetConstants(FRAGMENTSHADER& ps)
// indirect incoming texture scales, update all! // indirect incoming texture scales, update all!
if (s_bIndTexScaleChanged) { if (s_bIndTexScaleChanged) {
// set as two sets of vec4s, each containing S and T of two ind stages.
float f[8]; float f[8];
for (u32 i = 0; i < bpmem.genMode.numindstages; ++i) { for (u32 i = 0; i < bpmem.genMode.numindstages; ++i) {
@ -320,8 +318,8 @@ void PixelShaderMngr::SetConstants(FRAGMENTSHADER& ps)
int texcoord = bpmem.tevindref.getTexCoord(i); int texcoord = bpmem.tevindref.getTexCoord(i);
TCoordInfo& tc = bpmem.texcoords[texcoord]; TCoordInfo& tc = bpmem.texcoords[texcoord];
f[2*i] = bpmem.texscale[i/2].getScaleS(i&1) * (float)(tc.s.scale_minus_1+1) / (float)(lastTexDims[srctexmap]&0xffff); f[2*i] = bpmem.texscale[i/2].getScaleS(i&1) * (float)(tc.s.scale_minus_1+1) / (float)(lastTexDims[srctexmap] & 0xffff);
f[2*i+1] = bpmem.texscale[i/2].getScaleT(i&1) * (float)(tc.t.scale_minus_1+1) / (float)((lastTexDims[srctexmap]>>16)&0xfff); f[2*i+1] = bpmem.texscale[i/2].getScaleT(i&1) * (float)(tc.t.scale_minus_1+1) / (float)((lastTexDims[srctexmap] >> 16) & 0xffff);
PRIM_LOG("tex indscale%d: %f %f\n", i, f[2*i], f[2*i+1]); PRIM_LOG("tex indscale%d: %f %f\n", i, f[2*i], f[2*i+1]);
} }
@ -340,7 +338,7 @@ void PixelShaderMngr::SetConstants(FRAGMENTSHADER& ps)
int scale = ((u32)bpmem.indmtx[i].col0.s0 << 0) | int scale = ((u32)bpmem.indmtx[i].col0.s0 << 0) |
((u32)bpmem.indmtx[i].col1.s1 << 2) | ((u32)bpmem.indmtx[i].col1.s1 << 2) |
((u32)bpmem.indmtx[i].col2.s2 << 4); ((u32)bpmem.indmtx[i].col2.s2 << 4);
float fscale = powf(2.0f,(float)(scale - 17)) / 1024.0f; float fscale = powf(2.0f, (float)(scale - 17)) / 1024.0f;
// xyz - static matrix // xyz - static matrix
//TODO w - dynamic matrix scale / 256...... somehow / 4 works better //TODO w - dynamic matrix scale / 256...... somehow / 4 works better
@ -361,7 +359,6 @@ void PixelShaderMngr::SetConstants(FRAGMENTSHADER& ps)
void PixelShaderMngr::SetPSTextureDims(int texid) void PixelShaderMngr::SetPSTextureDims(int texid)
{ {
float fdims[4]; float fdims[4];
if (s_texturemask & (1<<texid)) { if (s_texturemask & (1<<texid)) {
if (maptocoord[texid] >= 0) { if (maptocoord[texid] >= 0) {
TCoordInfo& tc = bpmem.texcoords[maptocoord[texid]]; TCoordInfo& tc = bpmem.texcoords[maptocoord[texid]];
@ -392,7 +389,6 @@ void PixelShaderMngr::SetPSTextureDims(int texid)
fdims[3] = 1.0f/(float)((lastTexDims[texid]>>16)&0xfff); fdims[3] = 1.0f/(float)((lastTexDims[texid]>>16)&0xfff);
} }
} }
PRIM_LOG("texdims%d: %f %f %f %f\n", texid, fdims[0], fdims[1], fdims[2], fdims[3]); PRIM_LOG("texdims%d: %f %f %f %f\n", texid, fdims[0], fdims[1], fdims[2], fdims[3]);
SetPSConstant4fv(C_TEXDIMS + texid, fdims); SetPSConstant4fv(C_TEXDIMS + texid, fdims);
} }
@ -414,26 +410,26 @@ void PixelShaderMngr::SetColorChanged(int type, int num)
void PixelShaderMngr::SetAlpha(const AlphaFunc& alpha) void PixelShaderMngr::SetAlpha(const AlphaFunc& alpha)
{ {
if ((alpha.hex&0xffff) != lastAlpha) { if ((alpha.hex & 0xffff) != lastAlpha) {
lastAlpha = (lastAlpha&~0xffff)|(alpha.hex&0xffff); lastAlpha = (lastAlpha & ~0xffff) | (alpha.hex & 0xffff);
s_bAlphaChanged = true; s_bAlphaChanged = true;
} }
} }
void PixelShaderMngr::SetDestAlpha(const ConstantAlpha& alpha) void PixelShaderMngr::SetDestAlpha(const ConstantAlpha& alpha)
{ {
if (alpha.alpha != (lastAlpha>>16)) { if (alpha.alpha != (lastAlpha >> 16)) {
lastAlpha = (lastAlpha&~0xff0000)|((alpha.hex&0xff)<<16); lastAlpha = (lastAlpha & ~0xff0000) | ((alpha.hex & 0xff) << 16);
s_bAlphaChanged = true; s_bAlphaChanged = true;
} }
} }
void PixelShaderMngr::SetTexDims(int texmapid, u32 width, u32 height, u32 wraps, u32 wrapt) void PixelShaderMngr::SetTexDims(int texmapid, u32 width, u32 height, u32 wraps, u32 wrapt)
{ {
u32 wh = width | (height<<16) | (wraps<<28) | (wrapt<<30); u32 wh = width | (height << 16) | (wraps << 28) | (wrapt << 30);
if (lastTexDims[texmapid] != wh) { if (lastTexDims[texmapid] != wh) {
lastTexDims[texmapid] = wh; lastTexDims[texmapid] = wh;
s_nTexDimsChanged |= 1<<texmapid; s_nTexDimsChanged |= 1 << texmapid;
} }
} }
@ -487,7 +483,7 @@ void PixelShaderMngr::SetTexturesUsed(u32 nonpow2tex)
if (nonpow2tex & (0x10101 << i)) { if (nonpow2tex & (0x10101 << i)) {
// this check was previously implicit, but should it be here? // this check was previously implicit, but should it be here?
if (s_nTexDimsChanged ) if (s_nTexDimsChanged )
s_nTexDimsChanged |= 1<<i; s_nTexDimsChanged |= 1 << i;
} }
} }
s_texturemask = nonpow2tex; s_texturemask = nonpow2tex;
@ -497,7 +493,7 @@ void PixelShaderMngr::SetTexturesUsed(u32 nonpow2tex)
void PixelShaderMngr::SetTexDimsChanged(int texmapid) void PixelShaderMngr::SetTexDimsChanged(int texmapid)
{ {
// this check was previously implicit, but should it be here? // this check was previously implicit, but should it be here?
if (s_nTexDimsChanged ) if (s_nTexDimsChanged)
s_nTexDimsChanged |= 1 << texmapid; s_nTexDimsChanged |= 1 << texmapid;
SetIndTexScaleChanged(); SetIndTexScaleChanged();
@ -517,6 +513,8 @@ GLuint PixelShaderMngr::GetColorMatrixProgram()
return s_ColorMatrixProgram; return s_ColorMatrixProgram;
} }
// Mash together all the inputs that contribute to the code of a generated pixel shader into
// a unique identifier, basically containing all the bits. Yup, it's a lot ....
void PixelShaderMngr::GetPixelShaderId(PixelShaderMngr::PIXELSHADERUID& uid) void PixelShaderMngr::GetPixelShaderId(PixelShaderMngr::PIXELSHADERUID& uid)
{ {
u32 projtexcoords = 0; u32 projtexcoords = 0;
@ -524,25 +522,28 @@ void PixelShaderMngr::GetPixelShaderId(PixelShaderMngr::PIXELSHADERUID& uid)
if (bpmem.tevorders[i/2].getEnable(i&1)) { if (bpmem.tevorders[i/2].getEnable(i&1)) {
int texcoord = bpmem.tevorders[i/2].getTexCoord(i&1); int texcoord = bpmem.tevorders[i/2].getTexCoord(i&1);
if (xfregs.texcoords[texcoord].texmtxinfo.projection ) if (xfregs.texcoords[texcoord].texmtxinfo.projection )
projtexcoords |= 1<<texcoord; projtexcoords |= 1 << texcoord;
} }
} }
u32 zbufrender = (Renderer::GetZBufferTarget() && bpmem.zmode.updateenable) ? 1 : 0;
u32 zBufRenderToCol0 = Renderer::GetRenderMode() != Renderer::RM_Normal;
uid.values[0] = (u32)bpmem.genMode.numtevstages |
((u32)bpmem.genMode.numindstages << 4) |
((u32)bpmem.genMode.numtexgens << 7) |
((u32)bpmem.dstalpha.enable << 11) |
((u32)((bpmem.alphaFunc.hex >> 16) & 0xff) << 12) |
(projtexcoords << 20) |
((u32)bpmem.ztex2.op << 28) |
(zbufrender << 30) |
(zBufRenderToCol0 << 31);
u32 zbufrender = (Renderer::GetZBufferTarget()&&bpmem.zmode.updateenable)?1:0; s_curuid.values[0] = (s_curuid.values[0] & ~0x0ff00000) | (projtexcoords << 20);
u32 zBufRenderToCol0 = Renderer::GetRenderMode()!=Renderer::RM_Normal;
uid.values[0] = (u32)bpmem.genMode.numtevstages|((u32)bpmem.genMode.numindstages<<4)|((u32)bpmem.genMode.numtexgens<<7)
|((u32)bpmem.dstalpha.enable<<11)|((u32)((bpmem.alphaFunc.hex>>16)&0xff)<<12)|(projtexcoords<<20)|((u32)bpmem.ztex2.op<<28)
|(zbufrender<<30)|(zBufRenderToCol0<<31);
s_curuid.values[0] = (s_curuid.values[0]&~0x0ff00000)|(projtexcoords<<20);
// swap table // swap table
for (int i = 0; i < 8; i += 2) for (int i = 0; i < 8; i += 2)
((u8*)&uid.values[1])[i/2] = (bpmem.tevksel[i].hex&0xf)|((bpmem.tevksel[i+1].hex&0xf)<<4); ((u8*)&uid.values[1])[i/2] = (bpmem.tevksel[i].hex & 0xf) | ((bpmem.tevksel[i + 1].hex & 0xf)<<4);
uid.values[2] = s_texturemask; uid.values[2] = s_texturemask;
int hdr = 3; int hdr = 3;
u32* pcurvalue = &uid.values[hdr]; u32* pcurvalue = &uid.values[hdr];
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages+1; ++i) { for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages+1; ++i) {
TevStageCombiner::ColorCombiner &cc = bpmem.combiners[i].colorC; TevStageCombiner::ColorCombiner &cc = bpmem.combiners[i].colorC;
@ -552,18 +553,21 @@ void PixelShaderMngr::GetPixelShaderId(PixelShaderMngr::PIXELSHADERUID& uid)
u32 val1 = ac.hex&0xffffff; u32 val1 = ac.hex&0xffffff;
val0 |= bpmem.tevksel[i/2].getKC(i&1)<<24; val0 |= bpmem.tevksel[i/2].getKC(i&1)<<24;
val1 |= bpmem.tevksel[i/2].getKA(i&1)<<24; val1 |= bpmem.tevksel[i/2].getKA(i&1)<<24;
pcurvalue[0] = val0; pcurvalue[0] = val0;
pcurvalue[1] = val1; pcurvalue[1] = val1;
pcurvalue+=2; pcurvalue += 2;
} }
for (u32 i = 0; i < ((u32)bpmem.genMode.numtevstages+1)/2; ++i) { for (u32 i = 0; i < ((u32)bpmem.genMode.numtevstages+1)/2; ++i) {
u32 val0, val1; u32 val0, val1;
if (bpmem.tevorders[i].hex&0x40 ) val0 = bpmem.tevorders[i].hex&0x3ff; if (bpmem.tevorders[i].hex & 0x40)
else val0 = bpmem.tevorders[i].hex&0x380; val0 = bpmem.tevorders[i].hex & 0x3ff;
if (bpmem.tevorders[i].hex&0x40000 ) val1 = (bpmem.tevorders[i].hex&0x3ff000)>>12; else
else val1 = (bpmem.tevorders[i].hex&0x380000)>>12; val0 = bpmem.tevorders[i].hex & 0x380;
if (bpmem.tevorders[i].hex & 0x40000)
val1 = (bpmem.tevorders[i].hex & 0x3ff000) >> 12;
else
val1 = (bpmem.tevorders[i].hex & 0x380000) >> 12;
switch (i % 3) { switch (i % 3) {
case 0: pcurvalue[0] = val0|(val1<<10); break; case 0: pcurvalue[0] = val0|(val1<<10); break;
@ -572,31 +576,34 @@ void PixelShaderMngr::GetPixelShaderId(PixelShaderMngr::PIXELSHADERUID& uid)
} }
} }
if ((bpmem.genMode.numtevstages+1)&1) { // odd if ((bpmem.genMode.numtevstages + 1) & 1) { // odd
u32 val0; u32 val0;
if (bpmem.tevorders[bpmem.genMode.numtevstages/2].hex&0x40 ) val0 = bpmem.tevorders[bpmem.genMode.numtevstages/2].hex&0x3ff; if (bpmem.tevorders[bpmem.genMode.numtevstages/2].hex & 0x40)
else val0 = bpmem.tevorders[bpmem.genMode.numtevstages/2].hex&0x380; val0 = bpmem.tevorders[bpmem.genMode.numtevstages/2].hex&0x3ff;
else
val0 = bpmem.tevorders[bpmem.genMode.numtevstages/2].hex & 0x380;
switch (bpmem.genMode.numtevstages % 3) { switch (bpmem.genMode.numtevstages % 3) {
case 0: pcurvalue[0] = val0; break; case 0: pcurvalue[0] = val0; break;
case 1: pcurvalue[0] |= val0<<20; break; case 1: pcurvalue[0] |= val0 << 20; break;
case 2: pcurvalue[1] |= (val0<<10); pcurvalue++; break; case 2: pcurvalue[1] |= val0 << 10; pcurvalue++; break;
} }
} }
if ((bpmem.genMode.numtevstages % 3) != 2 ) if ((bpmem.genMode.numtevstages % 3) != 2)
++pcurvalue; ++pcurvalue;
uid.tevstages = (u32)(pcurvalue-&uid.values[0]-hdr); uid.tevstages = (u32)(pcurvalue-&uid.values[0]-hdr);
for (u32 i = 0; i < bpmem.genMode.numindstages; ++i) { for (u32 i = 0; i < bpmem.genMode.numindstages; ++i) {
u32 val = bpmem.tevind[i].hex&0x1fffff; // 21 bits u32 val = bpmem.tevind[i].hex & 0x1fffff; // 21 bits
switch (i%3) { switch (i%3) {
case 0: pcurvalue[0] = val; break; case 0: pcurvalue[0] = val; break;
case 1: pcurvalue[0] |= val<<21; pcurvalue[1] = val>>11; ++pcurvalue; break; case 1: pcurvalue[0] |= val << 21; pcurvalue[1] = val >> 11; ++pcurvalue; break;
case 2: pcurvalue[0] |= val<<10; ++pcurvalue; break; case 2: pcurvalue[0] |= val << 10; ++pcurvalue; break;
} }
} }
uid.indstages = (u32)(pcurvalue-&uid.values[0]-2-uid.tevstages); // yeah, well ....
uid.indstages = (u32)(pcurvalue - &uid.values[0] - 2 - uid.tevstages);
} }

80
Source/Plugins/Plugin_VideoOGL/Src/VertexLoader.cpp
View File

@ -137,13 +137,15 @@ VertexLoader::~VertexLoader()
int VertexLoader::ComputeVertexSize() int VertexLoader::ComputeVertexSize()
{ {
if (!m_AttrDirty) { if (!m_AttrDirty) {
if (m_VtxDesc.Hex0 == VertexManager::GetVtxDesc().Hex0 && (m_VtxDesc.Hex1&1)==(VertexManager::GetVtxDesc().Hex1&1)) // Compare the 33 desc bits.
if (m_VtxDesc.Hex0 == VertexManager::GetVtxDesc().Hex0 &&
(m_VtxDesc.Hex1 & 1) == (VertexManager::GetVtxDesc().Hex1 & 1))
return m_VertexSize; return m_VertexSize;
m_VtxDesc.Hex = VertexManager::GetVtxDesc().Hex; m_VtxDesc.Hex = VertexManager::GetVtxDesc().Hex;
} }
else { else {
// set anyway // Attributes are dirty so we have to recompute everything anyway.
m_VtxDesc.Hex = VertexManager::GetVtxDesc().Hex; m_VtxDesc.Hex = VertexManager::GetVtxDesc().Hex;
} }
@ -157,14 +159,14 @@ int VertexLoader::ComputeVertexSize()
m_VertexSize += 1; m_VertexSize += 1;
// Texture matrix indices // Texture matrix indices
if (m_VtxDesc.Tex0MatIdx) {m_VertexSize+=1; } if (m_VtxDesc.Tex0MatIdx) m_VertexSize += 1;
if (m_VtxDesc.Tex1MatIdx) {m_VertexSize+=1; } if (m_VtxDesc.Tex1MatIdx) m_VertexSize += 1;
if (m_VtxDesc.Tex2MatIdx) {m_VertexSize+=1; } if (m_VtxDesc.Tex2MatIdx) m_VertexSize += 1;
if (m_VtxDesc.Tex3MatIdx) {m_VertexSize+=1; } if (m_VtxDesc.Tex3MatIdx) m_VertexSize += 1;
if (m_VtxDesc.Tex4MatIdx) {m_VertexSize+=1; } if (m_VtxDesc.Tex4MatIdx) m_VertexSize += 1;
if (m_VtxDesc.Tex5MatIdx) {m_VertexSize+=1; } if (m_VtxDesc.Tex5MatIdx) m_VertexSize += 1;
if (m_VtxDesc.Tex6MatIdx) {m_VertexSize+=1; } if (m_VtxDesc.Tex6MatIdx) m_VertexSize += 1;
if (m_VtxDesc.Tex7MatIdx) {m_VertexSize+=1; } if (m_VtxDesc.Tex7MatIdx) m_VertexSize += 1;
switch (m_VtxDesc.Position) { switch (m_VtxDesc.Position) {
case NOT_PRESENT: {_assert_("Vertex descriptor without position!");} break; case NOT_PRESENT: {_assert_("Vertex descriptor without position!");} break;
@ -181,10 +183,10 @@ int VertexLoader::ComputeVertexSize()
} }
break; break;
case INDEX8: case INDEX8:
m_VertexSize+=1; m_VertexSize += 1;
break; break;
case INDEX16: case INDEX16:
m_VertexSize+=2; m_VertexSize += 2;
break; break;
} }
@ -202,20 +204,20 @@ int VertexLoader::ComputeVertexSize()
case DIRECT: case DIRECT:
switch (m_VtxAttr.color[i].Comp) switch (m_VtxAttr.color[i].Comp)
{ {
case FORMAT_16B_565: m_VertexSize+=2; break; case FORMAT_16B_565: m_VertexSize += 2; break;
case FORMAT_24B_888: m_VertexSize+=3; break; case FORMAT_24B_888: m_VertexSize += 3; break;
case FORMAT_32B_888x: m_VertexSize+=4; break; case FORMAT_32B_888x: m_VertexSize += 4; break;
case FORMAT_16B_4444: m_VertexSize+=2; break; case FORMAT_16B_4444: m_VertexSize += 2; break;
case FORMAT_24B_6666: m_VertexSize+=3; break; case FORMAT_24B_6666: m_VertexSize += 3; break;
case FORMAT_32B_8888: m_VertexSize+=4; break; case FORMAT_32B_8888: m_VertexSize += 4; break;
default: _assert_(0); break; default: _assert_(0); break;
} }
break; break;
case INDEX8: case INDEX8:
m_VertexSize+=1; m_VertexSize += 1;
break; break;
case INDEX16: case INDEX16:
m_VertexSize+=2; m_VertexSize += 2;
break; break;
} }
} }
@ -244,10 +246,10 @@ int VertexLoader::ComputeVertexSize()
} }
break; break;
case INDEX8: case INDEX8:
m_VertexSize+=1; m_VertexSize += 1;
break; break;
case INDEX16: case INDEX16:
m_VertexSize+=2; m_VertexSize += 2;
break; break;
} }
} }
@ -317,32 +319,32 @@ void VertexLoader::ProcessFormat()
case DIRECT: case DIRECT:
{ {
switch (m_VtxAttr.PosFormat) { switch (m_VtxAttr.PosFormat) {
case FORMAT_UBYTE: WriteCall(Pos_ReadDirect_UByte); break; case FORMAT_UBYTE: WriteCall(Pos_ReadDirect_UByte); break;
case FORMAT_BYTE: WriteCall(Pos_ReadDirect_Byte); break; case FORMAT_BYTE: WriteCall(Pos_ReadDirect_Byte); break;
case FORMAT_USHORT: WriteCall(Pos_ReadDirect_UShort); break; case FORMAT_USHORT: WriteCall(Pos_ReadDirect_UShort); break;
case FORMAT_SHORT: WriteCall(Pos_ReadDirect_Short); break; case FORMAT_SHORT: WriteCall(Pos_ReadDirect_Short); break;
case FORMAT_FLOAT: WriteCall(Pos_ReadDirect_Float); break; case FORMAT_FLOAT: WriteCall(Pos_ReadDirect_Float); break;
default: _assert_(0); break; default: _assert_(0); break;
} }
} }
break; break;
case INDEX8: case INDEX8:
switch (m_VtxAttr.PosFormat) { switch (m_VtxAttr.PosFormat) {
case FORMAT_UBYTE: WriteCall(Pos_ReadIndex8_UByte); break; //WTF? case FORMAT_UBYTE: WriteCall(Pos_ReadIndex8_UByte); break; //WTF?
case FORMAT_BYTE: WriteCall(Pos_ReadIndex8_Byte); break; case FORMAT_BYTE: WriteCall(Pos_ReadIndex8_Byte); break;
case FORMAT_USHORT: WriteCall(Pos_ReadIndex8_UShort); break; case FORMAT_USHORT: WriteCall(Pos_ReadIndex8_UShort); break;
case FORMAT_SHORT: WriteCall(Pos_ReadIndex8_Short); break; case FORMAT_SHORT: WriteCall(Pos_ReadIndex8_Short); break;
case FORMAT_FLOAT: WriteCall(Pos_ReadIndex8_Float); break; case FORMAT_FLOAT: WriteCall(Pos_ReadIndex8_Float); break;
default: _assert_(0); break; default: _assert_(0); break;
} }
break; break;
case INDEX16: case INDEX16:
switch (m_VtxAttr.PosFormat) { switch (m_VtxAttr.PosFormat) {
case FORMAT_UBYTE: WriteCall(Pos_ReadIndex16_UByte); break; case FORMAT_UBYTE: WriteCall(Pos_ReadIndex16_UByte); break;
case FORMAT_BYTE: WriteCall(Pos_ReadIndex16_Byte); break; case FORMAT_BYTE: WriteCall(Pos_ReadIndex16_Byte); break;
case FORMAT_USHORT: WriteCall(Pos_ReadIndex16_UShort); break; case FORMAT_USHORT: WriteCall(Pos_ReadIndex16_UShort); break;
case FORMAT_SHORT: WriteCall(Pos_ReadIndex16_Short); break; case FORMAT_SHORT: WriteCall(Pos_ReadIndex16_Short); break;
case FORMAT_FLOAT: WriteCall(Pos_ReadIndex16_Float); break; case FORMAT_FLOAT: WriteCall(Pos_ReadIndex16_Float); break;
default: _assert_(0); break; default: _assert_(0); break;
} }
break; break;
@ -396,8 +398,7 @@ void VertexLoader::ProcessFormat()
// Texture matrix indices (remove if corresponding texture coordinate isn't enabled) // Texture matrix indices (remove if corresponding texture coordinate isn't enabled)
for (int i = 0; i < 8; i++) { for (int i = 0; i < 8; i++) {
SetupTexCoord(i, tc[i], m_VtxAttr.texCoord[i].Format, m_VtxAttr.texCoord[i].Elements, m_VtxAttr.texCoord[i].Frac); SetupTexCoord(i, tc[i], m_VtxAttr.texCoord[i].Format, m_VtxAttr.texCoord[i].Elements, m_VtxAttr.texCoord[i].Frac);
if (m_components & (VB_HAS_TEXMTXIDX0 << i)) {
if (m_components&(VB_HAS_TEXMTXIDX0<<i)) {
if (tc[i] != NOT_PRESENT) { if (tc[i] != NOT_PRESENT) {
// if texmtx is included, texcoord will always be 3 floats, z will be the texmtx index // if texmtx is included, texcoord will always be 3 floats, z will be the texmtx index
WriteCall(m_VtxAttr.texCoord[i].Elements ? TexMtx_Write_Float : TexMtx_Write_Float2); WriteCall(m_VtxAttr.texCoord[i].Elements ? TexMtx_Write_Float : TexMtx_Write_Float2);
@ -412,7 +413,7 @@ void VertexLoader::ProcessFormat()
} }
else { else {
if (tc[i] != NOT_PRESENT) if (tc[i] != NOT_PRESENT)
m_VBVertexStride += 4 * (m_VtxAttr.texCoord[i].Elements?2:1); m_VBVertexStride += 4 * (m_VtxAttr.texCoord[i].Elements ? 2 : 1);
} }
if (tc[i] == NOT_PRESENT) { if (tc[i] == NOT_PRESENT) {
@ -719,9 +720,10 @@ void VertexLoader::RunVertices(int primitive, int count)
if ((v % granularity) == 0) if ((v % granularity) == 0)
{ {
if (VertexManager::GetRemainingSize() < granularity*m_VBVertexStride) { if (VertexManager::GetRemainingSize() < granularity*m_VBVertexStride) {
// This buffer full - break current primitive and flush, to switch to the next buffer.
u8* plastptr = VertexManager::s_pCurBufferPointer; u8* plastptr = VertexManager::s_pCurBufferPointer;
if (v - startv > 0) if (v - startv > 0)
VertexManager::AddVertices(primitive, v-startv+extraverts); VertexManager::AddVertices(primitive, v - startv + extraverts);
VertexManager::Flush(); VertexManager::Flush();
// Why does this need to be so complicated? // Why does this need to be so complicated?
switch (primitive) { switch (primitive) {

21
Source/Plugins/Plugin_VideoOGL/Src/VertexManager.cpp
View File

@ -24,7 +24,7 @@ static u32 s_prevcomponents; // previous state set
u8* VertexManager::s_pCurBufferPointer = NULL; u8* VertexManager::s_pCurBufferPointer = NULL;
const GLenum c_primitiveType[8] = static const GLenum c_primitiveType[8] =
{ {
GL_QUADS, GL_QUADS,
0, //nothing 0, //nothing
@ -57,6 +57,7 @@ bool VertexManager::Init()
glEnableClientState(GL_VERTEX_ARRAY); glEnableClientState(GL_VERTEX_ARRAY);
fnSetupVertexPointers = NULL; fnSetupVertexPointers = NULL;
s_vStoredPrimitives.reserve(1000);
GL_REPORT_ERRORD(); GL_REPORT_ERRORD();
return true; return true;
@ -149,7 +150,7 @@ void VertexManager::Flush()
GL_REPORT_ERRORD(); GL_REPORT_ERRORD();
glBindBuffer(GL_ARRAY_BUFFER, s_vboBuffers[s_nCurVBOIndex]); glBindBuffer(GL_ARRAY_BUFFER, s_vboBuffers[s_nCurVBOIndex]);
glBufferData(GL_ARRAY_BUFFER, s_pCurBufferPointer-s_pBaseBufferPointer, s_pBaseBufferPointer, GL_STREAM_DRAW); glBufferData(GL_ARRAY_BUFFER, s_pCurBufferPointer - s_pBaseBufferPointer, s_pBaseBufferPointer, GL_STREAM_DRAW);
GL_REPORT_ERRORD(); GL_REPORT_ERRORD();
// setup the pointers // setup the pointers
@ -159,17 +160,17 @@ void VertexManager::Flush()
// set the textures // set the textures
{ {
DVProfileFunc _pf("VertexManager::Flush:textures"); DVProfileFunc _pf("VertexManager::Flush:textures");
u32 usedtextures = 0; u32 usedtextures = 0;
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i) { for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i) {
if (bpmem.tevorders[i/2].getEnable(i & 1)) if (bpmem.tevorders[i/2].getEnable(i & 1))
usedtextures |= 1<<bpmem.tevorders[i/2].getTexMap(i & 1); usedtextures |= 1 << bpmem.tevorders[i/2].getTexMap(i & 1);
} }
if (bpmem.genMode.numindstages > 0) { if (bpmem.genMode.numindstages > 0) {
for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i) { for (u32 i = 0; i < (u32)bpmem.genMode.numtevstages + 1; ++i) {
if (bpmem.tevind[i].IsActive() && bpmem.tevind[i].bt < bpmem.genMode.numindstages) { if (bpmem.tevind[i].IsActive() && bpmem.tevind[i].bt < bpmem.genMode.numindstages) {
usedtextures |= 1<<bpmem.tevindref.getTexMap(bpmem.tevind[i].bt); usedtextures |= 1 << bpmem.tevindref.getTexMap(bpmem.tevind[i].bt);
} }
} }
} }
@ -177,7 +178,7 @@ void VertexManager::Flush()
u32 nonpow2tex = 0; u32 nonpow2tex = 0;
for (int i = 0; i < 8; i++) { for (int i = 0; i < 8; i++) {
if (usedtextures & (1 << i)) { if (usedtextures & (1 << i)) {
glActiveTexture(GL_TEXTURE0+i); glActiveTexture(GL_TEXTURE0 + i);
FourTexUnits &tex = bpmem.tex[i>>2]; FourTexUnits &tex = bpmem.tex[i>>2];
TextureMngr::TCacheEntry* tentry = TextureMngr::Load(i, (tex.texImage3[i&3].image_base/* & 0x1FFFFF*/) << 5, TextureMngr::TCacheEntry* tentry = TextureMngr::Load(i, (tex.texImage3[i&3].image_base/* & 0x1FFFFF*/) << 5,
@ -245,6 +246,9 @@ void VertexManager::Flush()
PixelShaderMngr::SetConstants(*ps); PixelShaderMngr::SetConstants(*ps);
// finally bind // finally bind
// TODO - cache progid, check if same as before. Maybe GL does this internally, though.
// This is the really annoying problem with GL - you never know whether it's worth caching stuff yourself.
glBindProgramARB(GL_VERTEX_PROGRAM_ARB, vs->glprogid); glBindProgramARB(GL_VERTEX_PROGRAM_ARB, vs->glprogid);
glBindProgramARB(GL_FRAGMENT_PROGRAM_ARB, ps->glprogid); glBindProgramARB(GL_FRAGMENT_PROGRAM_ARB, ps->glprogid);
@ -253,7 +257,8 @@ void VertexManager::Flush()
#endif #endif
int offset = 0; int offset = 0;
for (vector< pair<int, int> >::const_iterator it = s_vStoredPrimitives.begin(); it != s_vStoredPrimitives.end(); ++it) { for (vector< pair<int, int> >::const_iterator it = s_vStoredPrimitives.begin(); it != s_vStoredPrimitives.end(); ++it)
{
glDrawArrays(it->first, offset, it->second); glDrawArrays(it->first, offset, it->second);
offset += it->second; offset += it->second;
} }
@ -314,7 +319,7 @@ void VertexManager::LoadCPReg(u32 SubCmd, u32 Value)
case 0x90: g_VertexLoaders[SubCmd & 7].SetVAT_group2(Value); _assert_((SubCmd & 0x0F) < 8); break; case 0x90: g_VertexLoaders[SubCmd & 7].SetVAT_group2(Value); _assert_((SubCmd & 0x0F) < 8); break;
case 0xA0: arraybases[SubCmd & 0xF] = Value & 0xFFFFFFFF; break; case 0xA0: arraybases[SubCmd & 0xF] = Value & 0xFFFFFFFF; break;
case 0xB0: arraystrides[SubCmd & 0xF] = Value & 0xFF; break; case 0xB0: arraystrides[SubCmd & 0xF] = Value & 0xFF; break;
} }
} }

181
Source/Plugins/Plugin_VideoOGL/Src/VertexShaderManager.cpp
View File

@ -36,24 +36,23 @@
VertexShaderMngr::VSCache VertexShaderMngr::vshaders; VertexShaderMngr::VSCache VertexShaderMngr::vshaders;
VERTEXSHADER* VertexShaderMngr::pShaderLast = NULL; VERTEXSHADER* VertexShaderMngr::pShaderLast = NULL;
float VertexShaderMngr::rawViewport[6] = {0};
float VertexShaderMngr::rawProjection[7] = {0};
float GC_ALIGNED16(g_fProjectionMatrix[16]); float GC_ALIGNED16(g_fProjectionMatrix[16]);
static int s_nMaxVertexInstructions;
extern int A, B; extern int A, B;
extern float AR; extern float AR;
extern int nBackbufferWidth, nBackbufferHeight; extern int nBackbufferWidth, nBackbufferHeight;
//////////////////////// // Internal Variables
// Internal Variables // static int s_nMaxVertexInstructions;
////////////////////////
static float s_fMaterials[16];
static float s_fMaterials[16];
static float rawViewport[6] = {0};
static float rawProjection[7] = {0};
// track changes // track changes
static bool bTexMatricesChanged[2], bPosNormalMatrixChanged, bProjectionChanged, bViewportChanged; static bool bTexMatricesChanged[2], bPosNormalMatrixChanged, bProjectionChanged, bViewportChanged;
int nMaterialsChanged; static int nMaterialsChanged;
static int nTransformMatricesChanged[2]; // min,max static int nTransformMatricesChanged[2]; // min,max
static int nNormalMatricesChanged[2]; // min,max static int nNormalMatricesChanged[2]; // min,max
static int nPostTransformMatricesChanged[2]; // min,max static int nPostTransformMatricesChanged[2]; // min,max
@ -85,12 +84,15 @@ void VertexShaderMngr::Init()
void VertexShaderMngr::Shutdown() void VertexShaderMngr::Shutdown()
{ {
VSCache::iterator iter = vshaders.begin(); for (VSCache::iterator iter = vshaders.begin(); iter != vshaders.end(); iter++)
for (;iter!=vshaders.end();iter++)
iter->second.Destroy(); iter->second.Destroy();
vshaders.clear(); vshaders.clear();
} }
float VertexShaderMngr::GetPixelAspectRatio() {
return rawViewport[0] != 0 ? (float)Renderer::GetTargetWidth() / 640.0f : 1.0f;
}
VERTEXSHADER* VertexShaderMngr::GetShader(u32 components) VERTEXSHADER* VertexShaderMngr::GetShader(u32 components)
{ {
DVSTARTPROFILE(); DVSTARTPROFILE();
@ -170,7 +172,7 @@ bool VertexShaderMngr::CompileVertexShader(VERTEXSHADER& vs, const char* pstrpro
{ {
char stropt[64]; char stropt[64];
sprintf(stropt, "MaxLocalParams=256,MaxInstructions=%d", s_nMaxVertexInstructions); sprintf(stropt, "MaxLocalParams=256,MaxInstructions=%d", s_nMaxVertexInstructions);
const char* opts[] = {"-profileopts",stropt,"-O2", "-q", NULL}; const char *opts[] = {"-profileopts", stropt, "-O2", "-q", NULL};
CGprogram tempprog = cgCreateProgram(g_cgcontext, CG_SOURCE, pstrprogram, g_cgvProf, "main", opts); CGprogram tempprog = cgCreateProgram(g_cgcontext, CG_SOURCE, pstrprogram, g_cgvProf, "main", opts);
if (!cgIsProgram(tempprog) || cgGetError() != CG_NO_ERROR) { if (!cgIsProgram(tempprog) || cgGetError() != CG_NO_ERROR) {
ERROR_LOG("Failed to load vs %s:\n", cgGetLastListing(g_cgcontext)); ERROR_LOG("Failed to load vs %s:\n", cgGetLastListing(g_cgcontext));
@ -190,9 +192,9 @@ bool VertexShaderMngr::CompileVertexShader(VERTEXSHADER& vs, const char* pstrpro
plocal = strstr(plocal+13, "program.local"); plocal = strstr(plocal+13, "program.local");
} }
glGenProgramsARB( 1, &vs.glprogid ); glGenProgramsARB(1, &vs.glprogid);
glBindProgramARB( GL_VERTEX_PROGRAM_ARB, vs.glprogid ); glBindProgramARB(GL_VERTEX_PROGRAM_ARB, vs.glprogid);
glProgramStringARB( GL_VERTEX_PROGRAM_ARB, GL_PROGRAM_FORMAT_ASCII_ARB, (GLsizei)strlen(pcompiledprog), pcompiledprog); glProgramStringARB(GL_VERTEX_PROGRAM_ARB, GL_PROGRAM_FORMAT_ASCII_ARB, (GLsizei)strlen(pcompiledprog), pcompiledprog);
GLenum err = GL_NO_ERROR; GLenum err = GL_NO_ERROR;
GL_REPORT_ERROR(); GL_REPORT_ERROR();
@ -254,25 +256,31 @@ void VertexShaderMngr::SetConstants(VERTEXSHADER& vs)
if (nPostTransformMatricesChanged[0] >= 0) { if (nPostTransformMatricesChanged[0] >= 0) {
int startn = nPostTransformMatricesChanged[0]/4; int startn = nPostTransformMatricesChanged[0]/4;
int endn = (nPostTransformMatricesChanged[1]+3)/4; int endn = (nPostTransformMatricesChanged[1]+3)/4;
const float* pstart = (const float*)&xfmem[XFMEM_POSTMATRICES+startn*4]; const float* pstart = (const float*)&xfmem[XFMEM_POSTMATRICES + startn*4];
for(int i = startn; i < endn; ++i, pstart += 4) for(int i = startn; i < endn; ++i, pstart += 4)
SetVSConstant4fv(C_POSTTRANSFORMMATRICES+i, pstart); SetVSConstant4fv(C_POSTTRANSFORMMATRICES + i, pstart);
} }
if (nLightsChanged[0] >= 0) { if (nLightsChanged[0] >= 0) {
// lights don't have a 1 to 1 mapping, the color component needs to be converted to 4 floats // lights don't have a 1 to 1 mapping, the color component needs to be converted to 4 floats
int istart = nLightsChanged[0]/0x10; int istart = nLightsChanged[0] / 0x10;
int iend = (nLightsChanged[1]+15)/0x10; int iend = (nLightsChanged[1] + 15) / 0x10;
const float* xfmemptr = (const float*)&xfmem[0x10*istart+XFMEM_LIGHTS]; const float* xfmemptr = (const float*)&xfmem[0x10*istart + XFMEM_LIGHTS];
for(int i = istart; i < iend; ++i) { for (int i = istart; i < iend; ++i) {
u32 color = *(const u32*)(xfmemptr+3); u32 color = *(const u32*)(xfmemptr + 3);
SetVSConstant4f(C_LIGHTS+5*i, SetVSConstant4f(C_LIGHTS + 5*i,
((color>>24)&0xFF)/255.0f, ((color>>16)&0xFF)/255.0f, ((color>>8)&0xFF)/255.0f, ((color)&0xFF)/255.0f); ((color >> 24) & 0xFF)/255.0f,
((color >> 16) & 0xFF)/255.0f,
((color >> 8) & 0xFF)/255.0f,
((color) & 0xFF)/255.0f);
xfmemptr += 4; xfmemptr += 4;
for(int j = 0; j < 4; ++j, xfmemptr += 3) { for (int j = 0; j < 4; ++j, xfmemptr += 3) {
if( j == 1 && fabs(xfmemptr[0]) < 0.00001f && fabs(xfmemptr[1]) < 0.00001f && fabs(xfmemptr[2]) < 0.00001f) { if (j == 1 &&
// dist atten, make sure not equal to 0!!! fabs(xfmemptr[0]) < 0.00001f &&
fabs(xfmemptr[1]) < 0.00001f &&
fabs(xfmemptr[2]) < 0.00001f) {
// dist attenuation, make sure not equal to 0!!!
SetVSConstant4f(C_LIGHTS+5*i+j+1, 0.00001f, xfmemptr[1], xfmemptr[2], 0); SetVSConstant4f(C_LIGHTS+5*i+j+1, 0.00001f, xfmemptr[1], xfmemptr[2], 0);
} }
else else
@ -284,9 +292,9 @@ void VertexShaderMngr::SetConstants(VERTEXSHADER& vs)
} }
if (nMaterialsChanged) { if (nMaterialsChanged) {
for(int i = 0; i < 4; ++i) { for (int i = 0; i < 4; ++i) {
if( nMaterialsChanged&(1<<i) ) if (nMaterialsChanged & (1 << i))
SetVSConstant4fv(C_MATERIALS+i, &s_fMaterials[4*i]); SetVSConstant4fv(C_MATERIALS + i, &s_fMaterials[4*i]);
} }
nMaterialsChanged = 0; nMaterialsChanged = 0;
} }
@ -307,11 +315,12 @@ void VertexShaderMngr::SetConstants(VERTEXSHADER& vs)
if (bTexMatricesChanged[0]) { if (bTexMatricesChanged[0]) {
bTexMatricesChanged[0] = false; bTexMatricesChanged[0] = false;
float* fptrs[] = {
float* fptrs[] = {(float*)xfmem + MatrixIndexA.Tex0MtxIdx * 4, (float*)xfmem + MatrixIndexA.Tex1MtxIdx * 4, (float*)xfmem + MatrixIndexA.Tex0MtxIdx * 4, (float*)xfmem + MatrixIndexA.Tex1MtxIdx * 4,
(float*)xfmem + MatrixIndexA.Tex2MtxIdx * 4, (float*)xfmem + MatrixIndexA.Tex3MtxIdx * 4 }; (float*)xfmem + MatrixIndexA.Tex2MtxIdx * 4, (float*)xfmem + MatrixIndexA.Tex3MtxIdx * 4
};
for(int i = 0; i < 4; ++i) { for (int i = 0; i < 4; ++i) {
SetVSConstant4fv(C_TEXMATRICES+3*i, fptrs[i]); SetVSConstant4fv(C_TEXMATRICES+3*i, fptrs[i]);
SetVSConstant4fv(C_TEXMATRICES+3*i+1, fptrs[i]+4); SetVSConstant4fv(C_TEXMATRICES+3*i+1, fptrs[i]+4);
SetVSConstant4fv(C_TEXMATRICES+3*i+2, fptrs[i]+8); SetVSConstant4fv(C_TEXMATRICES+3*i+2, fptrs[i]+8);
@ -324,7 +333,7 @@ void VertexShaderMngr::SetConstants(VERTEXSHADER& vs)
float* fptrs[] = {(float*)xfmem + MatrixIndexB.Tex4MtxIdx * 4, (float*)xfmem + MatrixIndexB.Tex5MtxIdx * 4, float* fptrs[] = {(float*)xfmem + MatrixIndexB.Tex4MtxIdx * 4, (float*)xfmem + MatrixIndexB.Tex5MtxIdx * 4,
(float*)xfmem + MatrixIndexB.Tex6MtxIdx * 4, (float*)xfmem + MatrixIndexB.Tex7MtxIdx * 4 }; (float*)xfmem + MatrixIndexB.Tex6MtxIdx * 4, (float*)xfmem + MatrixIndexB.Tex7MtxIdx * 4 };
for(int i = 0; i < 4; ++i) { for (int i = 0; i < 4; ++i) {
SetVSConstant4fv(C_TEXMATRICES+3*i+12, fptrs[i]); SetVSConstant4fv(C_TEXMATRICES+3*i+12, fptrs[i]);
SetVSConstant4fv(C_TEXMATRICES+3*i+12+1, fptrs[i]+4); SetVSConstant4fv(C_TEXMATRICES+3*i+12+1, fptrs[i]+4);
SetVSConstant4fv(C_TEXMATRICES+3*i+12+2, fptrs[i]+8); SetVSConstant4fv(C_TEXMATRICES+3*i+12+2, fptrs[i]+8);
@ -350,10 +359,12 @@ void VertexShaderMngr::SetConstants(VERTEXSHADER& vs)
// rawViewport[0] = 320, rawViewport[1] = -240 // rawViewport[0] = 320, rawViewport[1] = -240
int scissorXOff = bpmem.scissorOffset.x * 2 - 342; int scissorXOff = bpmem.scissorOffset.x * 2 - 342;
int scissorYOff = bpmem.scissorOffset.y * 2 - 342; int scissorYOff = bpmem.scissorOffset.y * 2 - 342;
float fourThree = 4.0f/3.0f; float fourThree = 4.0f / 3.0f;
float ratio = AR / fourThree; float ratio = AR / fourThree;
float hAdj; float wAdj; float actualRatiow; float actualRatioh; float wAdj, hAdj;
int overfl; int xoffs = 0; int yoffs = 0; float actualRatiow, actualRatioh;
int overfl;
int xoffs = 0, yoffs = 0;
int wid, hei, actualWid, actualHei; int wid, hei, actualWid, actualHei;
int winw = nBackbufferWidth; int winh = nBackbufferHeight; int winw = nBackbufferWidth; int winh = nBackbufferHeight;
if (g_Config.bKeepAR) if (g_Config.bKeepAR)
@ -396,18 +407,18 @@ void VertexShaderMngr::SetConstants(VERTEXSHADER& vs)
if(g_Config.bStretchToFit && g_Config.renderToMainframe) if(g_Config.bStretchToFit && g_Config.renderToMainframe)
{ {
glViewport( glViewport(
(int)(rawViewport[3]-rawViewport[0]-342-scissorXOff) + xoffs, (int)(rawViewport[3]-rawViewport[0]-342-scissorXOff) + xoffs,
Renderer::GetTargetHeight() - ((int)(rawViewport[4]-rawViewport[1]-342-scissorYOff)) + yoffs, Renderer::GetTargetHeight() - ((int)(rawViewport[4]-rawViewport[1]-342-scissorYOff)) + yoffs,
wid, // width wid, // width
hei // height hei // height
); );
} }
else else
{ {
glViewport((int)(rawViewport[3]-rawViewport[0]-342-scissorXOff) * MValueX, glViewport((int)(rawViewport[3]-rawViewport[0]-342-scissorXOff) * MValueX,
Renderer::GetTargetHeight()-((int)(rawViewport[4]-rawViewport[1]-342-scissorYOff)) * MValueY, Renderer::GetTargetHeight()-((int)(rawViewport[4]-rawViewport[1]-342-scissorYOff)) * MValueY,
abs((int)(2 * rawViewport[0])) * MValueX, abs((int)(2 * rawViewport[1])) * MValueY); abs((int)(2 * rawViewport[0])) * MValueX, abs((int)(2 * rawViewport[1])) * MValueY);
} }
glDepthRange(-(0.0f - (rawViewport[5]-rawViewport[2])/-16777215.0f), rawViewport[5]/16777215.0f); glDepthRange(-(0.0f - (rawViewport[5]-rawViewport[2])/-16777215.0f), rawViewport[5]/16777215.0f);
@ -420,17 +431,17 @@ void VertexShaderMngr::SetConstants(VERTEXSHADER& vs)
g_fProjectionMatrix[0] = rawProjection[0]; g_fProjectionMatrix[0] = rawProjection[0];
g_fProjectionMatrix[1] = 0.0f; g_fProjectionMatrix[1] = 0.0f;
g_fProjectionMatrix[2] = rawProjection[1]; g_fProjectionMatrix[2] = rawProjection[1];
g_fProjectionMatrix[3] = 0;//-0.5f/Renderer::GetTargetWidth(); g_fProjectionMatrix[3] = 0;
g_fProjectionMatrix[4] = 0.0f; g_fProjectionMatrix[4] = 0.0f;
g_fProjectionMatrix[5] = rawProjection[2]; g_fProjectionMatrix[5] = rawProjection[2];
g_fProjectionMatrix[6] = rawProjection[3]; g_fProjectionMatrix[6] = rawProjection[3];
g_fProjectionMatrix[7] = 0;//+0.5f/Renderer::GetTargetHeight(); g_fProjectionMatrix[7] = 0;
g_fProjectionMatrix[8] = 0.0f; g_fProjectionMatrix[8] = 0.0f;
g_fProjectionMatrix[9] = 0.0f; g_fProjectionMatrix[9] = 0.0f;
g_fProjectionMatrix[10] = rawProjection[4]; g_fProjectionMatrix[10] = rawProjection[4];
g_fProjectionMatrix[11] = -(0.0f - rawProjection[5]); g_fProjectionMatrix[11] = -(0.0f - rawProjection[5]); // Yes, it's important that it's done this way.
g_fProjectionMatrix[12] = 0.0f; g_fProjectionMatrix[12] = 0.0f;
g_fProjectionMatrix[13] = 0.0f; g_fProjectionMatrix[13] = 0.0f;
@ -451,7 +462,7 @@ void VertexShaderMngr::SetConstants(VERTEXSHADER& vs)
g_fProjectionMatrix[8] = 0.0f; g_fProjectionMatrix[8] = 0.0f;
g_fProjectionMatrix[9] = 0.0f; g_fProjectionMatrix[9] = 0.0f;
g_fProjectionMatrix[10] = rawProjection[4]; g_fProjectionMatrix[10] = rawProjection[4];
g_fProjectionMatrix[11] = -(0.0f - rawProjection[5]); g_fProjectionMatrix[11] = -(0.0f - rawProjection[5]); // Yes, it's important that it's done this way.
g_fProjectionMatrix[12] = 0; g_fProjectionMatrix[12] = 0;
g_fProjectionMatrix[13] = 0; g_fProjectionMatrix[13] = 0;
@ -564,10 +575,10 @@ void VertexShaderMngr::SetTexMatrixChangedB(u32 Value)
void VertexShaderMngr::SetViewport(float* _Viewport) void VertexShaderMngr::SetViewport(float* _Viewport)
{ {
// check for paper mario // Workaround for paper mario, yep this is bizarre.
for (size_t i = 0; i < ARRAYSIZE(rawViewport); ++i) { for (size_t i = 0; i < ARRAYSIZE(rawViewport); ++i) {
if( *(u32*)(_Viewport + i) == 0x7f800000 ) if (*(u32*)(_Viewport + i) == 0x7f800000) // invalid fp number
return; // invalid number return;
} }
memcpy(rawViewport, _Viewport, sizeof(rawViewport)); memcpy(rawViewport, _Viewport, sizeof(rawViewport));
bViewportChanged = true; bViewportChanged = true;
@ -589,7 +600,7 @@ void VertexShaderMngr::LoadXFReg(u32 transferSize, u32 baseAddress, u32 *pData)
{ {
u32 address = baseAddress; u32 address = baseAddress;
for (int i=0; i<(int)transferSize; i++) for (int i = 0; i < (int)transferSize; i++)
{ {
address = baseAddress + i; address = baseAddress + i;
@ -723,16 +734,16 @@ void VertexShaderMngr::LoadXFReg(u32 transferSize, u32 baseAddress, u32 *pData)
break; break;
case 0x1018: case 0x1018:
//_assert_msg_(GX_XF, 0, "XF matrixindex0"); //_assert_msg_(GX_XF, 0, "XF matrixindex0");
VertexShaderMngr::SetTexMatrixChangedA(data); //? SetTexMatrixChangedA(data); //?
break; break;
case 0x1019: case 0x1019:
//_assert_msg_(GX_XF, 0, "XF matrixindex1"); //_assert_msg_(GX_XF, 0, "XF matrixindex1");
VertexShaderMngr::SetTexMatrixChangedB(data); //? SetTexMatrixChangedB(data); //?
break; break;
case 0x101a: case 0x101a:
VertexManager::Flush(); VertexManager::Flush();
VertexShaderMngr::SetViewport((float*)&pData[i]); SetViewport((float*)&pData[i]);
i += 6; i += 6;
break; break;
@ -795,12 +806,12 @@ void VertexShaderMngr::LoadXFReg(u32 transferSize, u32 baseAddress, u32 *pData)
} }
} }
// Check docs for this sucker... // TODO - verify that it is correct. Seems to work, though.
void VertexShaderMngr::LoadIndexedXF(u32 val, int array) void VertexShaderMngr::LoadIndexedXF(u32 val, int array)
{ {
int index = val>>16; int index = val >> 16;
int address = val&0xFFF; //check mask int address = val & 0xFFF; //check mask
int size = ((val>>12)&0xF)+1; int size = ((val >> 12) & 0xF) + 1;
//load stuff from array to address in xf mem //load stuff from array to address in xf mem
VertexManager::Flush(); VertexManager::Flush();
@ -816,39 +827,49 @@ float* VertexShaderMngr::GetPosNormalMat()
return (float*)xfmem + MatrixIndexA.PosNormalMtxIdx * 4; return (float*)xfmem + MatrixIndexA.PosNormalMtxIdx * 4;
} }
// Mash together all the inputs that contribute to the code of a generated vertex shader into
// a unique identifier, basically containing all the bits. Yup, it's a lot ....
void VertexShaderMngr::GetVertexShaderId(VERTEXSHADERUID& id, u32 components) void VertexShaderMngr::GetVertexShaderId(VERTEXSHADERUID& id, u32 components)
{ {
u32 zbufrender = (bpmem.ztex2.op==ZTEXTURE_ADD)||Renderer::GetZBufferTarget()!=0; u32 zbufrender = (bpmem.ztex2.op == ZTEXTURE_ADD) || Renderer::GetZBufferTarget() != 0;
id.values[0] = components|(xfregs.numTexGens<<23)|(xfregs.nNumChans<<27)|((u32)xfregs.bEnableDualTexTransform<<29)|(zbufrender<<30); id.values[0] = components |
(xfregs.numTexGens << 23) |
(xfregs.nNumChans << 27) |
((u32)xfregs.bEnableDualTexTransform << 29) |
(zbufrender << 30);
for(int i = 0; i < 2; ++i) { for (int i = 0; i < 2; ++i) {
id.values[1+i] = xfregs.colChans[i].color.enablelighting?(u32)xfregs.colChans[i].color.hex:(u32)xfregs.colChans[i].color.matsource; id.values[1+i] = xfregs.colChans[i].color.enablelighting ?
id.values[1+i] |= (xfregs.colChans[i].alpha.enablelighting?(u32)xfregs.colChans[i].alpha.hex:(u32)xfregs.colChans[i].alpha.matsource)<<15; (u32)xfregs.colChans[i].color.hex :
(u32)xfregs.colChans[i].color.matsource;
id.values[1+i] |= (xfregs.colChans[i].alpha.enablelighting ?
(u32)xfregs.colChans[i].alpha.hex :
(u32)xfregs.colChans[i].alpha.matsource) << 15;
} }
// fog
id.values[1] |= (((u32)bpmem.fog.c_proj_fsel.fsel&3)<<30); // fog
id.values[2] |= (((u32)bpmem.fog.c_proj_fsel.fsel>>2)<<30); id.values[1] |= (((u32)bpmem.fog.c_proj_fsel.fsel & 3) << 30);
id.values[2] |= (((u32)bpmem.fog.c_proj_fsel.fsel >> 2) << 30);
u32* pcurvalue = &id.values[3]; u32* pcurvalue = &id.values[3];
for (int i = 0; i < xfregs.numTexGens; ++i) {
for(int i = 0; i < xfregs.numTexGens; ++i) {
TexMtxInfo tinfo = xfregs.texcoords[i].texmtxinfo; TexMtxInfo tinfo = xfregs.texcoords[i].texmtxinfo;
if( tinfo.texgentype != XF_TEXGEN_EMBOSS_MAP ) if (tinfo.texgentype != XF_TEXGEN_EMBOSS_MAP)
tinfo.hex &= 0x7ff; tinfo.hex &= 0x7ff;
if( tinfo.texgentype != XF_TEXGEN_REGULAR ) if (tinfo.texgentype != XF_TEXGEN_REGULAR)
tinfo.projection = 0; tinfo.projection = 0;
u32 val = ((tinfo.hex>>1)&0x1ffff); u32 val = ((tinfo.hex >> 1) & 0x1ffff);
if( xfregs.bEnableDualTexTransform && tinfo.texgentype == XF_TEXGEN_REGULAR ) { if (xfregs.bEnableDualTexTransform && tinfo.texgentype == XF_TEXGEN_REGULAR) {
// rewrite normalization and post index // rewrite normalization and post index
val |= ((u32)xfregs.texcoords[i].postmtxinfo.index<<17)|((u32)xfregs.texcoords[i].postmtxinfo.normalize<<23); val |= ((u32)xfregs.texcoords[i].postmtxinfo.index << 17) | ((u32)xfregs.texcoords[i].postmtxinfo.normalize << 23);
} }
switch(i & 3) { switch (i & 3) {
case 0: pcurvalue[0] |= val; break; case 0: pcurvalue[0] |= val; break;
case 1: pcurvalue[0] |= val<<24; pcurvalue[1] = val>>8; ++pcurvalue; break; case 1: pcurvalue[0] |= val << 24; pcurvalue[1] = val >> 8; ++pcurvalue; break;
case 2: pcurvalue[0] |= val<<16; pcurvalue[1] = val>>16; ++pcurvalue; break; case 2: pcurvalue[0] |= val << 16; pcurvalue[1] = val >> 16; ++pcurvalue; break;
case 3: pcurvalue[0] |= val<<8; ++pcurvalue; break; case 3: pcurvalue[0] |= val << 8; ++pcurvalue; break;
} }
} }
} }

4
Source/Plugins/Plugin_VideoOGL/Src/VertexShaderManager.h
View File

@ -123,9 +123,7 @@ public:
static void LoadIndexedXF(u32 val, int array); static void LoadIndexedXF(u32 val, int array);
static float* GetPosNormalMat(); static float* GetPosNormalMat();
static float GetPixelAspectRatio();
static float rawViewport[6];
static float rawProjection[7];
}; };