Torque3D/Engine/source/gfx/gl/gfxGLTextureObject.cpp
Areloch f66a14c6e1 Merge pull request #1599 from JeffProgrammer/glprofiling
Added missing OpenGL profiling blocks.
2016-05-24 01:26:39 -05:00

352 lines
12 KiB
C++

//-----------------------------------------------------------------------------
// Copyright (c) 2012 GarageGames, LLC
//
// Permission is hereby granted, free of charge, to any person obtaining a copy
// of this software and associated documentation files (the "Software"), to
// deal in the Software without restriction, including without limitation the
// rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
// sell copies of the Software, and to permit persons to whom the Software is
// furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
// IN THE SOFTWARE.
//-----------------------------------------------------------------------------
#include "console/console.h"
#include "gfx/gl/tGL/tGL.h"
#include "math/mRect.h"
#include "gfx/gl/gfxGLTextureObject.h"
#include "gfx/gfxDevice.h"
#include "gfx/gl/gfxGLEnumTranslate.h"
#include "gfx/gl/gfxGLTextureManager.h"
#include "gfx/gl/gfxGLUtils.h"
#include "gfx/gfxCardProfile.h"
GFXGLTextureObject::GFXGLTextureObject(GFXDevice * aDevice, GFXTextureProfile *profile) :
GFXTextureObject(aDevice, profile),
mBinding(GL_TEXTURE_2D),
mBytesPerTexel(4),
mLockedRectRect(0, 0, 0, 0),
mGLDevice(static_cast<GFXGLDevice*>(mDevice)),
mZombieCache(NULL),
mNeedInitSamplerState(true),
mFrameAllocatorMark(0),
mFrameAllocatorPtr(NULL)
{
AssertFatal(dynamic_cast<GFXGLDevice*>(mDevice), "GFXGLTextureObject::GFXGLTextureObject - Invalid device type, expected GFXGLDevice!");
glGenTextures(1, &mHandle);
glGenBuffers(1, &mBuffer);
}
GFXGLTextureObject::~GFXGLTextureObject()
{
glDeleteTextures(1, &mHandle);
glDeleteBuffers(1, &mBuffer);
delete[] mZombieCache;
kill();
}
GFXLockedRect* GFXGLTextureObject::lock(U32 mipLevel, RectI *inRect)
{
AssertFatal(mBinding != GL_TEXTURE_3D, "GFXGLTextureObject::lock - We don't support locking 3D textures yet");
U32 width = mTextureSize.x >> mipLevel;
U32 height = mTextureSize.y >> mipLevel;
if(inRect)
{
if((inRect->point.x + inRect->extent.x > width) || (inRect->point.y + inRect->extent.y > height))
AssertFatal(false, "GFXGLTextureObject::lock - Rectangle too big!");
mLockedRectRect = *inRect;
}
else
{
mLockedRectRect = RectI(0, 0, width, height);
}
mLockedRect.pitch = mLockedRectRect.extent.x * mBytesPerTexel;
// CodeReview [ags 12/19/07] This one texel boundary is necessary to keep the clipmap code from crashing. Figure out why.
U32 size = (mLockedRectRect.extent.x + 1) * (mLockedRectRect.extent.y + 1) * mBytesPerTexel;
AssertFatal(!mFrameAllocatorMark && !mFrameAllocatorPtr, "");
mFrameAllocatorMark = FrameAllocator::getWaterMark();
mFrameAllocatorPtr = (U8*)FrameAllocator::alloc( size );
mLockedRect.bits = mFrameAllocatorPtr;
#if TORQUE_DEBUG
mFrameAllocatorMarkGuard = FrameAllocator::getWaterMark();
#endif
if( !mLockedRect.bits )
return NULL;
return &mLockedRect;
}
void GFXGLTextureObject::unlock(U32 mipLevel)
{
if(!mLockedRect.bits)
return;
// I know this is in unlock, but in GL we actually do our submission in unlock.
PROFILE_SCOPE(GFXGLTextureObject_lockRT);
PRESERVE_TEXTURE(mBinding);
glBindTexture(mBinding, mHandle);
glBindBuffer(GL_PIXEL_UNPACK_BUFFER_ARB, mBuffer);
glBufferData(GL_PIXEL_UNPACK_BUFFER_ARB, (mLockedRectRect.extent.x + 1) * (mLockedRectRect.extent.y + 1) * mBytesPerTexel, mFrameAllocatorPtr, GL_STREAM_DRAW);
if(mBinding == GL_TEXTURE_2D)
glTexSubImage2D(mBinding, mipLevel, mLockedRectRect.point.x, mLockedRectRect.point.y,
mLockedRectRect.extent.x, mLockedRectRect.extent.y, GFXGLTextureFormat[mFormat], GFXGLTextureType[mFormat], NULL);
else if(mBinding == GL_TEXTURE_1D)
glTexSubImage1D(mBinding, mipLevel, (mLockedRectRect.point.x > 1 ? mLockedRectRect.point.x : mLockedRectRect.point.y),
(mLockedRectRect.extent.x > 1 ? mLockedRectRect.extent.x : mLockedRectRect.extent.y), GFXGLTextureFormat[mFormat], GFXGLTextureType[mFormat], NULL);
glBindBuffer(GL_PIXEL_UNPACK_BUFFER_ARB, 0);
mLockedRect.bits = NULL;
#if TORQUE_DEBUG
AssertFatal(mFrameAllocatorMarkGuard == FrameAllocator::getWaterMark(), "");
#endif
FrameAllocator::setWaterMark(mFrameAllocatorMark);
mFrameAllocatorMark = 0;
mFrameAllocatorPtr = NULL;
}
void GFXGLTextureObject::release()
{
glDeleteTextures(1, &mHandle);
glDeleteBuffers(1, &mBuffer);
mHandle = 0;
mBuffer = 0;
}
void GFXGLTextureObject::reInit()
{
AssertFatal(!mHandle && !mBuffer,"Must release before reInit");
glGenTextures(1, &mHandle);
glGenBuffers(1, &mBuffer);
}
bool GFXGLTextureObject::copyToBmp(GBitmap * bmp)
{
if (!bmp)
return false;
// check format limitations
// at the moment we only support RGBA for the source (other 4 byte formats should
// be easy to add though)
AssertFatal(mFormat == GFXFormatR8G8B8A8, "GFXGLTextureObject::copyToBmp - invalid format");
AssertFatal(bmp->getFormat() == GFXFormatR8G8B8A8 || bmp->getFormat() == GFXFormatR8G8B8, "GFXGLTextureObject::copyToBmp - invalid format");
if(mFormat != GFXFormatR8G8B8A8)
return false;
if(bmp->getFormat() != GFXFormatR8G8B8A8 && bmp->getFormat() != GFXFormatR8G8B8)
return false;
AssertFatal(bmp->getWidth() == getWidth(), "GFXGLTextureObject::copyToBmp - invalid size");
AssertFatal(bmp->getHeight() == getHeight(), "GFXGLTextureObject::copyToBmp - invalid size");
PROFILE_SCOPE(GFXGLTextureObject_copyToBmp);
PRESERVE_TEXTURE(mBinding);
glBindTexture(mBinding, mHandle);
U8 dstBytesPerPixel = GFXFormat_getByteSize( bmp->getFormat() );
U8 srcBytesPerPixel = GFXFormat_getByteSize( mFormat );
if(dstBytesPerPixel == srcBytesPerPixel)
{
glGetTexImage(mBinding, 0, GFXGLTextureFormat[mFormat], GFXGLTextureType[mFormat], bmp->getWritableBits());
return true;
}
FrameAllocatorMarker mem;
U32 srcPixelCount = mTextureSize.x * mTextureSize.y;
U8 *dest = bmp->getWritableBits();
U8 *orig = (U8*)mem.alloc(srcPixelCount * srcBytesPerPixel);
glGetTexImage(mBinding, 0, GFXGLTextureFormat[mFormat], GFXGLTextureType[mFormat], orig);
PROFILE_START(GFXGLTextureObject_copyToBmp_pixCopy);
for(int i = 0; i < srcPixelCount; ++i)
{
dest[0] = orig[0];
dest[1] = orig[1];
dest[2] = orig[2];
if(dstBytesPerPixel == 4)
dest[3] = orig[3];
orig += srcBytesPerPixel;
dest += dstBytesPerPixel;
}
PROFILE_END();
return true;
}
void GFXGLTextureObject::initSamplerState(const GFXSamplerStateDesc &ssd)
{
glTexParameteri(mBinding, GL_TEXTURE_MIN_FILTER, minificationFilter(ssd.minFilter, ssd.mipFilter, mMipLevels));
glTexParameteri(mBinding, GL_TEXTURE_MAG_FILTER, GFXGLTextureFilter[ssd.magFilter]);
glTexParameteri(mBinding, GL_TEXTURE_WRAP_S, !mIsNPoT2 ? GFXGLTextureAddress[ssd.addressModeU] : GL_CLAMP_TO_EDGE);
glTexParameteri(mBinding, GL_TEXTURE_WRAP_T, !mIsNPoT2 ? GFXGLTextureAddress[ssd.addressModeV] : GL_CLAMP_TO_EDGE);
if(mBinding == GL_TEXTURE_3D)
glTexParameteri(mBinding, GL_TEXTURE_WRAP_R, GFXGLTextureAddress[ssd.addressModeW]);
if(static_cast< GFXGLDevice* >( GFX )->supportsAnisotropic() )
glTexParameterf(mBinding, GL_TEXTURE_MAX_ANISOTROPY_EXT, ssd.maxAnisotropy);
mNeedInitSamplerState = false;
mSampler = ssd;
}
void GFXGLTextureObject::bind(U32 textureUnit)
{
glActiveTexture(GL_TEXTURE0 + textureUnit);
glBindTexture(mBinding, mHandle);
GFXGL->getOpenglCache()->setCacheBindedTex(textureUnit, mBinding, mHandle);
if(GFXGL->mCapabilities.samplerObjects)
return;
GFXGLStateBlockRef sb = mGLDevice->getCurrentStateBlock();
AssertFatal(sb, "GFXGLTextureObject::bind - No active stateblock!");
if (!sb)
return;
const GFXSamplerStateDesc ssd = sb->getDesc().samplers[textureUnit];
if(mNeedInitSamplerState)
{
initSamplerState(ssd);
return;
}
if(mSampler.minFilter != ssd.minFilter || mSampler.mipFilter != ssd.mipFilter)
glTexParameteri(mBinding, GL_TEXTURE_MIN_FILTER, minificationFilter(ssd.minFilter, ssd.mipFilter, mMipLevels));
if(mSampler.magFilter != ssd.magFilter)
glTexParameteri(mBinding, GL_TEXTURE_MAG_FILTER, GFXGLTextureFilter[ssd.magFilter]);
if(mSampler.addressModeU != ssd.addressModeU)
glTexParameteri(mBinding, GL_TEXTURE_WRAP_S, !mIsNPoT2 ? GFXGLTextureAddress[ssd.addressModeU] : GL_CLAMP_TO_EDGE);
if(mSampler.addressModeV != ssd.addressModeV)
glTexParameteri(mBinding, GL_TEXTURE_WRAP_T, !mIsNPoT2 ? GFXGLTextureAddress[ssd.addressModeV] : GL_CLAMP_TO_EDGE);
if(mBinding == GL_TEXTURE_3D && mSampler.addressModeW != ssd.addressModeW )
glTexParameteri(mBinding, GL_TEXTURE_WRAP_R, GFXGLTextureAddress[ssd.addressModeW]);
if(mSampler.maxAnisotropy != ssd.maxAnisotropy && static_cast< GFXGLDevice* >( GFX )->supportsAnisotropic() )
glTexParameterf(mBinding, GL_TEXTURE_MAX_ANISOTROPY_EXT, ssd.maxAnisotropy);
mSampler = ssd;
}
U8* GFXGLTextureObject::getTextureData( U32 mip )
{
AssertFatal( mMipLevels, "");
mip = (mip < mMipLevels) ? mip : 0;
const U32 dataSize = isCompressedFormat(mFormat)
? getCompressedSurfaceSize( mFormat, mTextureSize.x, mTextureSize.y, mip )
: (mTextureSize.x >> mip) * (mTextureSize.y >> mip) * mBytesPerTexel;
U8* data = new U8[dataSize];
PRESERVE_TEXTURE(mBinding);
glBindTexture(mBinding, mHandle);
if( isCompressedFormat(mFormat) )
glGetCompressedTexImage( mBinding, mip, data );
else
glGetTexImage(mBinding, mip, GFXGLTextureFormat[mFormat], GFXGLTextureType[mFormat], data);
return data;
}
void GFXGLTextureObject::copyIntoCache()
{
PRESERVE_TEXTURE(mBinding);
glBindTexture(mBinding, mHandle);
U32 cacheSize = mTextureSize.x * mTextureSize.y;
if(mBinding == GL_TEXTURE_3D)
cacheSize *= mTextureSize.z;
cacheSize *= mBytesPerTexel;
mZombieCache = new U8[cacheSize];
glGetTexImage(mBinding, 0, GFXGLTextureFormat[mFormat], GFXGLTextureType[mFormat], mZombieCache);
}
void GFXGLTextureObject::reloadFromCache()
{
if(!mZombieCache)
return;
if(mBinding == GL_TEXTURE_3D)
{
static_cast<GFXGLTextureManager*>(TEXMGR)->_loadTexture(this, mZombieCache);
delete[] mZombieCache;
mZombieCache = NULL;
return;
}
PRESERVE_TEXTURE(mBinding);
glBindTexture(mBinding, mHandle);
if(mBinding == GL_TEXTURE_2D)
glTexSubImage2D(mBinding, 0, 0, 0, mTextureSize.x, mTextureSize.y, GFXGLTextureFormat[mFormat], GFXGLTextureType[mFormat], mZombieCache);
else if(mBinding == GL_TEXTURE_1D)
glTexSubImage1D(mBinding, 0, 0, (mTextureSize.x > 1 ? mTextureSize.x : mTextureSize.y), GFXGLTextureFormat[mFormat], GFXGLTextureType[mFormat], mZombieCache);
if(mMipLevels != 1)
glGenerateMipmap(mBinding);
delete[] mZombieCache;
mZombieCache = NULL;
mIsZombie = false;
}
void GFXGLTextureObject::zombify()
{
if(mIsZombie)
return;
mIsZombie = true;
if(!mProfile->doStoreBitmap() && !mProfile->isRenderTarget() && !mProfile->isDynamic() && !mProfile->isZTarget())
copyIntoCache();
release();
}
void GFXGLTextureObject::resurrect()
{
if(!mIsZombie)
return;
glGenTextures(1, &mHandle);
glGenBuffers(1, &mBuffer);
}
F32 GFXGLTextureObject::getMaxUCoord() const
{
return mBinding == GL_TEXTURE_2D ? 1.0f : (F32)getWidth();
}
F32 GFXGLTextureObject::getMaxVCoord() const
{
return mBinding == GL_TEXTURE_2D ? 1.0f : (F32)getHeight();
}
const String GFXGLTextureObject::describeSelf() const
{
String ret = Parent::describeSelf();
ret += String::ToString(" GL Handle: %i", mHandle);
return ret;
}