refactor and stabilise texture cache

This commit is contained in:
RSDuck 2020-04-23 00:41:07 +02:00
parent c70519070b
commit bb1c7d8460
5 changed files with 137 additions and 228 deletions

View File

@ -33,6 +33,8 @@ int Threaded3D;
int GL_ScaleFactor; int GL_ScaleFactor;
int GL_Antialias; int GL_Antialias;
int GL_TextureFiltering;
int GL_TextureUpscaleFactor;
ConfigEntry ConfigFile[] = ConfigEntry ConfigFile[] =
{ {
@ -41,6 +43,8 @@ ConfigEntry ConfigFile[] =
{"GL_ScaleFactor", 0, &GL_ScaleFactor, 1, NULL, 0}, {"GL_ScaleFactor", 0, &GL_ScaleFactor, 1, NULL, 0},
{"GL_Antialias", 0, &GL_Antialias, 0, NULL, 0}, {"GL_Antialias", 0, &GL_Antialias, 0, NULL, 0},
{"GL_TextureFiltering", 0, &GL_TextureFiltering, 0, NULL, 0},
{"GL_TextureUpscaleFactor", 0, &GL_TextureUpscaleFactor, 1, NULL, 0},
{"", -1, NULL, 0, NULL, 0} {"", -1, NULL, 0, NULL, 0}
}; };

View File

@ -45,6 +45,8 @@ extern int Threaded3D;
extern int GL_ScaleFactor; extern int GL_ScaleFactor;
extern int GL_Antialias; extern int GL_Antialias;
extern int GL_TextureFiltering;
extern int GL_TextureUpscaleFactor;
} }

View File

@ -122,6 +122,8 @@ int FrontBuffer;
GLuint FramebufferID[4], PixelbufferID; GLuint FramebufferID[4], PixelbufferID;
u32 Framebuffer[256*192]; u32 Framebuffer[256*192];
u32 ConversionBuffer[1024*1024];
struct TextureAllocator struct TextureAllocator
{ {
u32 Length = 0; u32 Length = 0;
@ -137,31 +139,31 @@ inline TextureAllocator& GetTextureAllocator(u32 width, u32 height)
return TextureMem[__builtin_ctz(width) - 3][__builtin_ctz(height) - 3]; return TextureMem[__builtin_ctz(width) - 3][__builtin_ctz(height) - 3];
} }
u32 ConversionBuffer[1024*1024];
u32* AllocateTexture(TexCache::ExternalTexHandle* handle, u32 width, u32 height) u32* AllocateTexture(TexCache::ExternalTexHandle* handle, u32 width, u32 height)
{ {
TextureAllocator& allocator = GetTextureAllocator(width, height); TextureAllocator& allocator = GetTextureAllocator(width, height);
//printf("allocating texture %d %d\n", width, height);
u32 scaledWidth = width * 2; u32 scaledWidth = width * Config::GL_TextureUpscaleFactor;
u32 scaledHeight = height * 2; u32 scaledHeight = height * Config::GL_TextureUpscaleFactor;
if (allocator.FreeEntriesLeft == 0) if (allocator.FreeEntriesLeft == 0)
{ {
u32 newLength = (allocator.Length * 3) / 2 + 8; u32 newLength = (26 - __builtin_ctz(width * height)) + (allocator.Length * 3) / 2;
if (newLength >= 64*8) if (newLength > 64*8)
newLength = 64*8;
if (newLength == allocator.Length)
abort(); abort();
if (allocator.Length == 0) if (allocator.Length == 0)
{ {
printf("created new texture\n");
glGenTextures(1, &allocator.TextureID); glGenTextures(1, &allocator.TextureID);
glBindTexture(GL_TEXTURE_2D_ARRAY, allocator.TextureID); glBindTexture(GL_TEXTURE_2D_ARRAY, allocator.TextureID);
glTexParameteri(GL_TEXTURE_2D_ARRAY, GL_TEXTURE_MIN_FILTER, GL_LINEAR); GLenum filter = Config::GL_TextureFiltering ? GL_LINEAR : GL_NEAREST;
glTexParameteri(GL_TEXTURE_2D_ARRAY, GL_TEXTURE_MAG_FILTER, GL_LINEAR); glTexParameteri(GL_TEXTURE_2D_ARRAY, GL_TEXTURE_MIN_FILTER, filter);
glTexParameteri(GL_TEXTURE_2D_ARRAY, GL_TEXTURE_MAG_FILTER, filter);
glTexParameteri(GL_TEXTURE_2D_ARRAY, GL_TEXTURE_WRAP_S, GL_MIRRORED_REPEAT); glTexParameteri(GL_TEXTURE_2D_ARRAY, GL_TEXTURE_WRAP_S, GL_MIRRORED_REPEAT);
glTexParameteri(GL_TEXTURE_2D_ARRAY, GL_TEXTURE_WRAP_T, GL_MIRRORED_REPEAT); glTexParameteri(GL_TEXTURE_2D_ARRAY, GL_TEXTURE_WRAP_T, GL_MIRRORED_REPEAT);
} }
@ -194,8 +196,10 @@ u32* AllocateTexture(TexCache::ExternalTexHandle* handle, u32 width, u32 height)
allocator.FreeEntries[i] |= 1ULL << freeIdx; allocator.FreeEntries[i] |= 1ULL << freeIdx;
*handle = i * 64 + freeIdx; *handle = i * 64 + freeIdx;
//return &allocator.Pixels[*handle * scaledWidth * scaledHeight]; if (Config::GL_TextureUpscaleFactor > 1)
return ConversionBuffer; return ConversionBuffer;
else
return &allocator.Pixels[*handle * scaledWidth * scaledHeight];
} }
} }
@ -208,17 +212,18 @@ void FreeTexture(TexCache::ExternalTexHandle handle, u32 width, u32 height)
TextureAllocator& allocator = GetTextureAllocator(width, height); TextureAllocator& allocator = GetTextureAllocator(width, height);
allocator.FreeEntriesLeft++; allocator.FreeEntriesLeft++;
allocator.FreeEntries[handle >> 6] &= ~(1 << (handle & 0x3F)); allocator.FreeEntries[handle >> 6] &= ~(1ULL << (handle & 0x3F));
} }
void FinaliseTexture(TexCache::ExternalTexHandle handle, u32 width, u32 height) void FinaliseTexture(TexCache::ExternalTexHandle handle, u32 width, u32 height)
{ {
TextureAllocator& allocator = GetTextureAllocator(width, height); TextureAllocator& allocator = GetTextureAllocator(width, height);
u32 scaledWidth = width * 2; u32 scaledWidth = width * Config::GL_TextureUpscaleFactor;
u32 scaledHeight = height * 2; u32 scaledHeight = height * Config::GL_TextureUpscaleFactor;
xbrz::scale(2, ConversionBuffer, &allocator.Pixels[handle * scaledWidth * scaledHeight], width, height, xbrz::ColorFormat::ARGB); if (Config::GL_TextureUpscaleFactor > 1)
xbrz::scale(Config::GL_TextureUpscaleFactor, ConversionBuffer, &allocator.Pixels[handle * scaledWidth * scaledHeight], width, height, xbrz::ColorFormat::ARGB);
// could still be improved // could still be improved
glBindTexture(GL_TEXTURE_2D_ARRAY, allocator.TextureID); glBindTexture(GL_TEXTURE_2D_ARRAY, allocator.TextureID);
@ -459,6 +464,18 @@ bool Init()
void DeInit() void DeInit()
{ {
for (int j = 0; j < 8; j++)
{
for (int i = 0; i < 8; i++)
{
if (TextureMem[j][i].TextureID)
{
glDeleteTextures(1, &TextureMem[j][i].TextureID);
delete[] TextureMem[j][i].Pixels;
}
}
}
glDeleteFramebuffers(4, &FramebufferID[0]); glDeleteFramebuffers(4, &FramebufferID[0]);
glDeleteTextures(8, &FramebufferTex[0]); glDeleteTextures(8, &FramebufferTex[0]);
@ -478,6 +495,14 @@ void DeInit()
void Reset() void Reset()
{ {
for (int j = 0; j < 8; j++)
{
for (int i = 0; i < 8; i++)
{
memset(TextureMem[j][i].FreeEntries, 0, sizeof(TextureMem[j][i].FreeEntries));
TextureMem[j][i].FreeEntriesLeft = TextureMem[j][i].Length;
}
}
} }
void UpdateDisplaySettings() void UpdateDisplaySettings()
@ -558,6 +583,20 @@ void UpdateDisplaySettings()
glBindBuffer(GL_PIXEL_PACK_BUFFER, PixelbufferID); glBindBuffer(GL_PIXEL_PACK_BUFFER, PixelbufferID);
glBufferData(GL_PIXEL_PACK_BUFFER, 256*192*4, NULL, GL_DYNAMIC_READ); glBufferData(GL_PIXEL_PACK_BUFFER, 256*192*4, NULL, GL_DYNAMIC_READ);
for (int j = 0; j < 8; j++)
{
for (int i = 0; i < 8; i++)
{
if (TextureMem[j][i].TextureID != 0)
{
glBindTexture(GL_TEXTURE_2D, TextureMem[j][i].TextureID);
GLenum filter = Config::GL_TextureFiltering ? GL_LINEAR : GL_NEAREST;
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, filter);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, filter);
}
}
}
//glLineWidth(scale); //glLineWidth(scale);
//glLineWidth(1.5); //glLineWidth(1.5);
} }

View File

@ -25,6 +25,8 @@
#include <SDL2/SDL.h> #include <SDL2/SDL.h>
#include <assert.h>
namespace GPU3D namespace GPU3D
{ {
namespace SoftRenderer namespace SoftRenderer
@ -77,6 +79,7 @@ struct TextureAllocator
u64 FreeEntries[8]; u64 FreeEntries[8];
u32 FreeEntriesLeft = 0; u32 FreeEntriesLeft = 0;
}; };
bool TextureRealloc;
// all sizes below 8*8 (log2(64)=6) can be ignored // all sizes below 8*8 (log2(64)=6) can be ignored
TextureAllocator TextureMem[14]; TextureAllocator TextureMem[14];
@ -91,7 +94,7 @@ u32* AllocateTexture(TexCache::ExternalTexHandle* handle, u32 width, u32 height)
if (allocator.FreeEntriesLeft == 0) if (allocator.FreeEntriesLeft == 0)
{ {
u32 newLength = (20 - (__builtin_ctz(width * height) - 6) + 4) + (allocator.Length * 2) / 3; u32 newLength = 20 - (__builtin_ctz(width * height) - 6) + (allocator.Length * 3) / 2;
// while it's theoretically possible to hit this limit // while it's theoretically possible to hit this limit
// other things will probably break before it // other things will probably break before it
@ -101,6 +104,7 @@ u32* AllocateTexture(TexCache::ExternalTexHandle* handle, u32 width, u32 height)
u32* newPixels = new u32[width * height * newLength]; u32* newPixels = new u32[width * height * newLength];
if (allocator.Length) if (allocator.Length)
{ {
TextureRealloc = true;
memcpy(newPixels, allocator.Pixels, allocator.Length * width * height * 4); memcpy(newPixels, allocator.Pixels, allocator.Length * width * height * 4);
delete[] allocator.Pixels; delete[] allocator.Pixels;
} }
@ -119,6 +123,8 @@ u32* AllocateTexture(TexCache::ExternalTexHandle* handle, u32 width, u32 height)
allocator.FreeEntries[i] |= 1ULL << freeIdx; allocator.FreeEntries[i] |= 1ULL << freeIdx;
*handle = (i * 64 + freeIdx) * width * height; *handle = (i * 64 + freeIdx) * width * height;
assert(i * 64 + freeIdx < allocator.Length);
return &allocator.Pixels[*handle]; return &allocator.Pixels[*handle];
} }
} }
@ -132,8 +138,9 @@ void FreeTexture(TexCache::ExternalTexHandle handle, u32 width, u32 height)
TextureAllocator& allocator = GetTextureAllocator(width, height); TextureAllocator& allocator = GetTextureAllocator(width, height);
handle /= width * height; handle /= width * height;
assert(allocator.FreeEntries[handle >> 6] & (1ULL << (handle & 0x3F)));
allocator.FreeEntriesLeft++; allocator.FreeEntriesLeft++;
allocator.FreeEntries[handle >> 6] &= ~(1 << (handle & 0x3F)); allocator.FreeEntries[handle >> 6] &= ~(1ULL << (handle & 0x3F));
} }
// threading // threading
@ -198,6 +205,12 @@ bool Init()
void DeInit() void DeInit()
{ {
for (int i = 0; i < 14; i++)
{
if (TextureMem[i].Pixels)
delete[] TextureMem[i].Pixels;
}
StopRenderThread(); StopRenderThread();
Platform::Semaphore_Free(Sema_RenderStart); Platform::Semaphore_Free(Sema_RenderStart);
@ -209,8 +222,13 @@ void Reset()
{ {
for (int i = 0; i < 14; i++) for (int i = 0; i < 14; i++)
{ {
if (TextureMem[i].Pixels)
{
delete[] TextureMem[i].Pixels;
TextureMem[i].Pixels = NULL;
}
memset(TextureMem[i].FreeEntries, 0, sizeof(TextureMem[i].FreeEntries)); memset(TextureMem[i].FreeEntries, 0, sizeof(TextureMem[i].FreeEntries));
TextureMem[i].FreeEntriesLeft = TextureMem[i].Length * 64; TextureMem[i].FreeEntriesLeft = 0;
} }
memset(ColorBuffer, 0, BufferSize * 2 * 4); memset(ColorBuffer, 0, BufferSize * 2 * 4);
@ -676,185 +694,6 @@ u32 TextureLookup(u32* texture, u32 texparam, u32 texpal, s16 s, s16 t)
} }
return texture[s + t * width]; return texture[s + t * width];
/*
u8 alpha0;
if (texparam & (1<<29)) alpha0 = 0;
else alpha0 = 31;
switch ((texparam >> 26) & 0x7)
{
case 1: // A3I5
{
vramaddr += ((t * width) + s);
u8 pixel = GPU::ReadVRAM_Texture<u8>(vramaddr);
texpal <<= 4;
*color = GPU::ReadVRAM_TexPal<u16>(texpal + ((pixel&0x1F)<<1));
*alpha = ((pixel >> 3) & 0x1C) + (pixel >> 6);
}
break;
case 2: // 4-color
{
vramaddr += (((t * width) + s) >> 2);
u8 pixel = GPU::ReadVRAM_Texture<u8>(vramaddr);
pixel >>= ((s & 0x3) << 1);
pixel &= 0x3;
texpal <<= 3;
*color = GPU::ReadVRAM_TexPal<u16>(texpal + (pixel<<1));
*alpha = (pixel==0) ? alpha0 : 31;
}
break;
case 3: // 16-color
{
vramaddr += (((t * width) + s) >> 1);
u8 pixel = GPU::ReadVRAM_Texture<u8>(vramaddr);
if (s & 0x1) pixel >>= 4;
else pixel &= 0xF;
texpal <<= 4;
*color = GPU::ReadVRAM_TexPal<u16>(texpal + (pixel<<1));
*alpha = (pixel==0) ? alpha0 : 31;
}
break;
case 4: // 256-color
{
vramaddr += ((t * width) + s);
u8 pixel = GPU::ReadVRAM_Texture<u8>(vramaddr);
texpal <<= 4;
*color = GPU::ReadVRAM_TexPal<u16>(texpal + (pixel<<1));
*alpha = (pixel==0) ? alpha0 : 31;
}
break;
case 5: // compressed
{
vramaddr += ((t & 0x3FC) * (width>>2)) + (s & 0x3FC);
vramaddr += (t & 0x3);
u32 slot1addr = 0x20000 + ((vramaddr & 0x1FFFC) >> 1);
if (vramaddr >= 0x40000)
slot1addr += 0x10000;
u8 val = GPU::ReadVRAM_Texture<u8>(vramaddr);
val >>= (2 * (s & 0x3));
u16 palinfo = GPU::ReadVRAM_Texture<u16>(slot1addr);
u32 paloffset = (palinfo & 0x3FFF) << 2;
texpal <<= 4;
switch (val & 0x3)
{
case 0:
*color = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset);
*alpha = 31;
break;
case 1:
*color = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset + 2);
*alpha = 31;
break;
case 2:
if ((palinfo >> 14) == 1)
{
u16 color0 = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset);
u16 color1 = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset + 2);
u32 r0 = color0 & 0x001F;
u32 g0 = color0 & 0x03E0;
u32 b0 = color0 & 0x7C00;
u32 r1 = color1 & 0x001F;
u32 g1 = color1 & 0x03E0;
u32 b1 = color1 & 0x7C00;
u32 r = (r0 + r1) >> 1;
u32 g = ((g0 + g1) >> 1) & 0x03E0;
u32 b = ((b0 + b1) >> 1) & 0x7C00;
*color = r | g | b;
}
else if ((palinfo >> 14) == 3)
{
u16 color0 = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset);
u16 color1 = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset + 2);
u32 r0 = color0 & 0x001F;
u32 g0 = color0 & 0x03E0;
u32 b0 = color0 & 0x7C00;
u32 r1 = color1 & 0x001F;
u32 g1 = color1 & 0x03E0;
u32 b1 = color1 & 0x7C00;
u32 r = (r0*5 + r1*3) >> 3;
u32 g = ((g0*5 + g1*3) >> 3) & 0x03E0;
u32 b = ((b0*5 + b1*3) >> 3) & 0x7C00;
*color = r | g | b;
}
else
*color = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset + 4);
*alpha = 31;
break;
case 3:
if ((palinfo >> 14) == 2)
{
*color = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset + 6);
*alpha = 31;
}
else if ((palinfo >> 14) == 3)
{
u16 color0 = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset);
u16 color1 = GPU::ReadVRAM_TexPal<u16>(texpal + paloffset + 2);
u32 r0 = color0 & 0x001F;
u32 g0 = color0 & 0x03E0;
u32 b0 = color0 & 0x7C00;
u32 r1 = color1 & 0x001F;
u32 g1 = color1 & 0x03E0;
u32 b1 = color1 & 0x7C00;
u32 r = (r0*3 + r1*5) >> 3;
u32 g = ((g0*3 + g1*5) >> 3) & 0x03E0;
u32 b = ((b0*3 + b1*5) >> 3) & 0x7C00;
*color = r | g | b;
*alpha = 31;
}
else
{
*color = 0;
*alpha = 0;
}
break;
}
}
break;
case 6: // A5I3
{
vramaddr += ((t * width) + s);
u8 pixel = GPU::ReadVRAM_Texture<u8>(vramaddr);
texpal <<= 4;
*color = GPU::ReadVRAM_TexPal<u16>(texpal + ((pixel&0x7)<<1));
*alpha = (pixel >> 3);
}
break;
case 7: // direct color
{
vramaddr += (((t * width) + s) << 1);
*color = GPU::ReadVRAM_Texture<u16>(vramaddr);
*alpha = (*color & 0x8000) ? 31 : 0;
}
break;
}*/
} }
// depth test is 'less or equal' instead of 'less than' under the following conditions: // depth test is 'less or equal' instead of 'less than' under the following conditions:
@ -973,9 +812,9 @@ u32 RenderPixel(Polygon* polygon, u8 vr, u8 vg, u8 vb, s16 s, s16 t, u32* textur
{ {
u32 tcolor = TextureLookup(texture, polygon->TexParam, polygon->TexPalette, s, t); u32 tcolor = TextureLookup(texture, polygon->TexParam, polygon->TexPalette, s, t);
u8 tr = tcolor & 0x3E; if (tr) tr++; u8 tr = tcolor & 0x3E;
u8 tg = (tcolor >> 8) & 0x3E; if (tg) tg++; u8 tg = (tcolor >> 8) & 0x3E;
u8 tb = (tcolor >> 16) & 0x3E; if (tb) tb++; u8 tb = (tcolor >> 16) & 0x3E;
u8 talpha = tcolor >> 24; u8 talpha = tcolor >> 24;
if (blendmode & 0x1) if (blendmode & 0x1)
@ -1136,25 +975,8 @@ void SetupPolygonRightEdge(RendererPolygon* rp, s32 y)
polygon->FinalW[rp->CurVR], polygon->FinalW[rp->NextVR], y); polygon->FinalW[rp->CurVR], polygon->FinalW[rp->NextVR], y);
} }
void SetupPolygon(RendererPolygon* rp, Polygon* polygon, RendererPolygon* lastRp) void SetupPolygon(RendererPolygon* rp, Polygon* polygon)
{ {
if (polygon->TexParam & 0x1C000000 && !polygon->IsShadowMask)
{
if (lastRp && lastRp->PolyData->TexParam == polygon->TexParam
&& lastRp->PolyData->TexPalette == polygon->TexPalette)
{
rp->TextureData = lastRp->TextureData;
}
else
{
TexCache::ExternalTexHandle handle =
TexCache::GetTexture<TexCache::outputFmt_RGB6A5>(polygon->TexParam, polygon->TexPalette);
u32 width = 8 << ((polygon->TexParam >> 20) & 0x7);
u32 height = 8 << ((polygon->TexParam >> 23) & 0x7);
rp->TextureData = &GetTextureAllocator(width, height).Pixels[handle];
}
}
u32 nverts = polygon->NumVertices; u32 nverts = polygon->NumVertices;
u32 vtop = polygon->VTop, vbot = polygon->VBottom; u32 vtop = polygon->VTop, vbot = polygon->VBottom;
@ -2149,17 +1971,55 @@ void ClearBuffers()
} }
} }
void SetupTexture(RendererPolygon* rp, RendererPolygon* lastRp)
{
Polygon* polygon = rp->PolyData;
Polygon* lastPolygon = lastRp ? lastRp->PolyData : NULL;
if (lastRp
&& !lastPolygon->IsShadowMask
&& polygon->TexParam == lastPolygon->TexParam
&& polygon->TexPalette == lastPolygon->TexPalette)
{
rp->TextureData = lastRp->TextureData;
}
else if (polygon->TexParam & 0x1C000000 && !polygon->IsShadowMask)
{
TexCache::ExternalTexHandle handle =
TexCache::GetTexture<TexCache::outputFmt_RGB6A5>(polygon->TexParam, polygon->TexPalette);
u32 width = 8 << ((polygon->TexParam >> 20) & 0x7);
u32 height = 8 << ((polygon->TexParam >> 23) & 0x7);
rp->TextureData = &GetTextureAllocator(width, height).Pixels[handle];
}
else
{
rp->TextureData = NULL;
}
}
void RenderPolygons(bool threaded, Polygon** polygons, int npolys) void RenderPolygons(bool threaded, Polygon** polygons, int npolys)
{ {
TexCache::UpdateTextures(); TexCache::UpdateTextures();
int j = 0; int j = 0;
{
RendererPolygon* lastRp = NULL;
for (int i = 0; i < npolys; i++) for (int i = 0; i < npolys; i++)
{ {
if (polygons[i]->Degenerate) continue; if (polygons[i]->Degenerate) continue;
SetupPolygon(&PolygonList[j], polygons[i], j > 0 ? &PolygonList[j - 1] : NULL); SetupPolygon(&PolygonList[j], polygons[i]);
SetupTexture(&PolygonList[j], j ? &PolygonList[j - 1] : NULL);
j++; j++;
} }
}
if (TextureRealloc)
{
// realloating the buffer invalidated all pointers to the previous buffer
for (int i = 0; i < j; i++)
{
SetupTexture(&PolygonList[i], i ? &PolygonList[i - 1] : NULL);
}
TextureRealloc = 0;
}
TexCache::SaveTextures(); TexCache::SaveTextures();
RenderScanline(0, j); RenderScanline(0, j);

View File

@ -35,9 +35,13 @@ inline u32 ConvertRGB5ToBGR8(u16 val)
} }
inline u32 ConvertRGB5ToRGB6(u16 val) inline u32 ConvertRGB5ToRGB6(u16 val)
{ {
return (((u32)val & 0x1F) << 1) u8 r = (val & 0x1F) << 1;
| (((u32)val & 0x3E0) << 4) u8 g = (val & 0x3E0) >> 4;
| (((u32)val & 0x7C00) << 7); u8 b = (val & 0x7C00) >> 9;
if (r) r++;
if (g) g++;
if (b) b++;
return (u32)r | ((u32)g << 8) | ((u32)b << 16);
} }
template <int outputFmt> template <int outputFmt>
@ -356,8 +360,8 @@ void UpdateTextures()
{ {
if (GPU::VRAMMap_TexPal[i] != PaletteMap[i]) if (GPU::VRAMMap_TexPal[i] != PaletteMap[i])
{ {
PaletteDirty[i >> 2] |= 0xFFFF << (i & 0x3) * 16; PaletteDirty[i >> 2] |= 0xFFFFULL << (i & 0x3) * 16;
PaletteCacheStatus[i >> 2] &= ~(0xFFFF << (i & 0x3) * 16); PaletteCacheStatus[i >> 2] &= ~(0xFFFFULL << (i & 0x3) * 16);
PaletteMap[i] = GPU::VRAMMap_TexPal[i]; PaletteMap[i] = GPU::VRAMMap_TexPal[i];
} }
else else