// Copyright 2016 The SwiftShader Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//    http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

// VertexDataManager.h: Defines the VertexDataManager, a class that
// runs the Buffer translation process.

#include "VertexDataManager.h"

#include "Buffer.h"
#include "Program.h"
#include "IndexDataManager.h"
#include "common/debug.h"

namespace
{
	enum {INITIAL_STREAM_BUFFER_SIZE = 1024 * 1024};
}

namespace es2
{

VertexDataManager::VertexDataManager(Context *context) : mContext(context)
{
	for(int i = 0; i < MAX_VERTEX_ATTRIBS; i++)
	{
		mDirtyCurrentValue[i] = true;
		mCurrentValueBuffer[i] = nullptr;
	}

	mStreamingBuffer = new StreamingVertexBuffer(INITIAL_STREAM_BUFFER_SIZE);

	if(!mStreamingBuffer)
	{
		ERR("Failed to allocate the streaming vertex buffer.");
	}
}

VertexDataManager::~VertexDataManager()
{
	delete mStreamingBuffer;

	for(int i = 0; i < MAX_VERTEX_ATTRIBS; i++)
	{
		delete mCurrentValueBuffer[i];
	}
}

unsigned int VertexDataManager::writeAttributeData(StreamingVertexBuffer *vertexBuffer, GLint start, GLsizei count, const VertexAttribute &attribute)
{
	Buffer *buffer = attribute.mBoundBuffer;

	int inputStride = attribute.stride();
	int elementSize = attribute.typeSize();
	unsigned int streamOffset = 0;

	char *output = nullptr;

	if(vertexBuffer)
	{
		output = (char*)vertexBuffer->map(attribute, attribute.typeSize() * count, &streamOffset);
	}

	if(!output)
	{
		ERR("Failed to map vertex buffer.");
		return ~0u;
	}

	const char *input = nullptr;

	if(buffer)
	{
		input = static_cast<const char*>(buffer->data()) + attribute.mOffset;
	}
	else
	{
		input = static_cast<const char*>(attribute.mPointer);
	}

	input += inputStride * start;

	if(inputStride == elementSize)
	{
		memcpy(output, input, count * inputStride);
	}
	else
	{
		for(int i = 0; i < count; i++)
		{
			memcpy(output, input, elementSize);
			output += elementSize;
			input += inputStride;
		}
	}

	vertexBuffer->unmap();

	return streamOffset;
}

GLenum VertexDataManager::prepareVertexData(GLint start, GLsizei count, TranslatedAttribute *translated, GLsizei instanceId)
{
	if(!mStreamingBuffer)
	{
		return GL_OUT_OF_MEMORY;
	}

	const VertexAttributeArray &attribs = mContext->getVertexArrayAttributes();
	const VertexAttributeArray &currentAttribs = mContext->getCurrentVertexAttributes();
	Program *program = mContext->getCurrentProgram();

	// Determine the required storage size per used buffer
	for(int i = 0; i < MAX_VERTEX_ATTRIBS; i++)
	{
		const VertexAttribute &attrib = attribs[i].mArrayEnabled ? attribs[i] : currentAttribs[i];

		if(program->getAttributeStream(i) != -1 && attrib.mArrayEnabled)
		{
			if(!attrib.mBoundBuffer)
			{
				const bool isInstanced = attrib.mDivisor > 0;
				mStreamingBuffer->addRequiredSpace(attrib.typeSize() * (isInstanced ? 1 : count));
			}
		}
	}

	mStreamingBuffer->reserveRequiredSpace();

	// Perform the vertex data translations
	for(int i = 0; i < MAX_VERTEX_ATTRIBS; i++)
	{
		if(program->getAttributeStream(i) != -1)
		{
			const VertexAttribute &attrib = attribs[i].mArrayEnabled ? attribs[i] : currentAttribs[i];

			if(attrib.mArrayEnabled)
			{
				const bool isInstanced = attrib.mDivisor > 0;

				// Instanced vertices do not apply the 'start' offset
				GLint firstVertexIndex = isInstanced ? instanceId / attrib.mDivisor : start;

				Buffer *buffer = attrib.mBoundBuffer;

				if((!buffer && attrib.mPointer == nullptr) || (buffer && !buffer->data()))
				{
					// This is an application error that would normally result in a crash, but we catch it and return an error
					ERR("An enabled vertex array has no buffer and no pointer.");
					return GL_INVALID_OPERATION;
				}

				sw::Resource *staticBuffer = buffer ? buffer->getResource() : nullptr;

				if(staticBuffer)
				{
					translated[i].vertexBuffer = staticBuffer;
					translated[i].offset = firstVertexIndex * attrib.stride() + static_cast<int>(attrib.mOffset);
					translated[i].stride = isInstanced ? 0 : attrib.stride();
				}
				else
				{
					unsigned int streamOffset = writeAttributeData(mStreamingBuffer, firstVertexIndex, isInstanced ? 1 : count, attrib);

					if(streamOffset == ~0u)
					{
						return GL_OUT_OF_MEMORY;
					}

					translated[i].vertexBuffer = mStreamingBuffer->getResource();
					translated[i].offset = streamOffset;
					translated[i].stride = isInstanced ? 0 : attrib.typeSize();
				}

				switch(attrib.mType)
				{
				case GL_BYTE:           translated[i].type = sw::STREAMTYPE_SBYTE;  break;
				case GL_UNSIGNED_BYTE:  translated[i].type = sw::STREAMTYPE_BYTE;   break;
				case GL_SHORT:          translated[i].type = sw::STREAMTYPE_SHORT;  break;
				case GL_UNSIGNED_SHORT: translated[i].type = sw::STREAMTYPE_USHORT; break;
				case GL_INT:            translated[i].type = sw::STREAMTYPE_INT;    break;
				case GL_UNSIGNED_INT:   translated[i].type = sw::STREAMTYPE_UINT;   break;
				case GL_FIXED:          translated[i].type = sw::STREAMTYPE_FIXED;  break;
				case GL_FLOAT:          translated[i].type = sw::STREAMTYPE_FLOAT;  break;
				case GL_HALF_FLOAT:     translated[i].type = sw::STREAMTYPE_HALF;   break;
				case GL_HALF_FLOAT_OES: translated[i].type = sw::STREAMTYPE_HALF;   break;
				case GL_INT_2_10_10_10_REV:          translated[i].type = sw::STREAMTYPE_2_10_10_10_INT;  break;
				case GL_UNSIGNED_INT_2_10_10_10_REV: translated[i].type = sw::STREAMTYPE_2_10_10_10_UINT; break;
				default: UNREACHABLE(attrib.mType); translated[i].type = sw::STREAMTYPE_FLOAT;  break;
				}

				translated[i].count = attrib.mSize;
				translated[i].normalized = attrib.mNormalized;
			}
			else
			{
				if(mDirtyCurrentValue[i])
				{
					delete mCurrentValueBuffer[i];
					mCurrentValueBuffer[i] = new ConstantVertexBuffer(attrib.getCurrentValueBitsAsFloat(0), attrib.getCurrentValueBitsAsFloat(1), attrib.getCurrentValueBitsAsFloat(2), attrib.getCurrentValueBitsAsFloat(3));
					mDirtyCurrentValue[i] = false;
				}

				translated[i].vertexBuffer = mCurrentValueBuffer[i]->getResource();

				switch(attrib.currentValueType())
				{
				case GL_INT:
					translated[i].type = sw::STREAMTYPE_INT;
					break;
				case GL_UNSIGNED_INT:
					translated[i].type = sw::STREAMTYPE_UINT;
					break;
				default:
					translated[i].type = sw::STREAMTYPE_FLOAT;
					break;
				}
				translated[i].count = 4;
				translated[i].stride = 0;
				translated[i].offset = 0;
				translated[i].normalized = false;
			}
		}
	}

	return GL_NO_ERROR;
}

VertexBuffer::VertexBuffer(unsigned int size) : mVertexBuffer(nullptr)
{
	if(size > 0)
	{
		mVertexBuffer = new sw::Resource(size + 1024);

		if(!mVertexBuffer)
		{
			ERR("Out of memory allocating a vertex buffer of size %u.", size);
		}
	}
}

VertexBuffer::~VertexBuffer()
{
	if(mVertexBuffer)
	{
		mVertexBuffer->destruct();
	}
}

void VertexBuffer::unmap()
{
	if(mVertexBuffer)
	{
		mVertexBuffer->unlock();
	}
}

sw::Resource *VertexBuffer::getResource() const
{
	return mVertexBuffer;
}

ConstantVertexBuffer::ConstantVertexBuffer(float x, float y, float z, float w) : VertexBuffer(4 * sizeof(float))
{
	if(mVertexBuffer)
	{
		float *vector = (float*)mVertexBuffer->lock(sw::PUBLIC);

		vector[0] = x;
		vector[1] = y;
		vector[2] = z;
		vector[3] = w;

		mVertexBuffer->unlock();
	}
}

ConstantVertexBuffer::~ConstantVertexBuffer()
{
}

StreamingVertexBuffer::StreamingVertexBuffer(unsigned int size) : VertexBuffer(size)
{
	mBufferSize = size;
	mWritePosition = 0;
	mRequiredSpace = 0;
}

StreamingVertexBuffer::~StreamingVertexBuffer()
{
}

void StreamingVertexBuffer::addRequiredSpace(unsigned int requiredSpace)
{
	mRequiredSpace += requiredSpace;
}

void *StreamingVertexBuffer::map(const VertexAttribute &attribute, unsigned int requiredSpace, unsigned int *offset)
{
	void *mapPtr = nullptr;

	if(mVertexBuffer)
	{
		// We can use a private lock because we never overwrite the content
		mapPtr = (char*)mVertexBuffer->lock(sw::PRIVATE) + mWritePosition;

		*offset = mWritePosition;
		mWritePosition += requiredSpace;
	}

	return mapPtr;
}

void StreamingVertexBuffer::reserveRequiredSpace()
{
	if(mRequiredSpace > mBufferSize)
	{
		if(mVertexBuffer)
		{
			mVertexBuffer->destruct();
			mVertexBuffer = 0;
		}

		mBufferSize = std::max(mRequiredSpace, 3 * mBufferSize / 2);   // 1.5 x mBufferSize is arbitrary and should be checked to see we don't have too many reallocations.

		mVertexBuffer = new sw::Resource(mBufferSize);

		if(!mVertexBuffer)
		{
			ERR("Out of memory allocating a vertex buffer of size %u.", mBufferSize);
		}

		mWritePosition = 0;
	}
	else if(mWritePosition + mRequiredSpace > mBufferSize)   // Recycle
	{
		if(mVertexBuffer)
		{
			mVertexBuffer->destruct();
			mVertexBuffer = new sw::Resource(mBufferSize);
		}

		mWritePosition = 0;
	}

	mRequiredSpace = 0;
}

}