/*
* Copyright (C) 2015-2016 Intel Corporation. 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.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "common/common_def.h"
#include "common/log.h"
#include "ocl/oclcontext.h"
#include "oclpostprocess_blender.h"
#include "oclvppimage.h"
namespace YamiMediaCodec{
YamiStatus
OclPostProcessBlender::blend(const SharedPtr<VideoFrame>& src,
const SharedPtr<VideoFrame>& dst)
{
YamiStatus ret = YAMI_SUCCESS;
SharedPtr<OclVppCLImage> srcImagePtr, dstImagePtr;
cl_image_format srcFormat, dstFormat;
srcFormat.image_channel_order = CL_RGBA;
srcFormat.image_channel_data_type = CL_UNORM_INT8;
srcImagePtr = OclVppCLImage::create(m_display, src, m_context, srcFormat);
if (!srcImagePtr) {
ERROR("failed to create cl image from src frame");
return YAMI_FAIL;
}
dstFormat.image_channel_order = CL_RG;
dstFormat.image_channel_data_type = CL_UNORM_INT8;
dstImagePtr = OclVppCLImage::create(m_display, dst, m_context, dstFormat);
if (!dstImagePtr) {
ERROR("failed to create cl image from dst frame");
return YAMI_FAIL;
}
cl_mem bgImageMem[3];
for (uint32_t n = 0; n < dstImagePtr->numPlanes(); n++) {
bgImageMem[n] = dstImagePtr->plane(n);
}
uint32_t pixelSize = getPixelSize(dstFormat);
VideoRect crop;
crop.x = dst->crop.x / pixelSize;
crop.y = dst->crop.y & ~1;
crop.width = dst->crop.width / pixelSize;
crop.height = dst->crop.height;
if ((CL_SUCCESS != clSetKernelArg(m_kernelBlend, 0, sizeof(cl_mem), &dstImagePtr->plane(0)))
|| (CL_SUCCESS != clSetKernelArg(m_kernelBlend, 1, sizeof(cl_mem), &dstImagePtr->plane(1)))
|| (CL_SUCCESS != clSetKernelArg(m_kernelBlend, 2, sizeof(cl_mem), &bgImageMem[0]))
|| (CL_SUCCESS != clSetKernelArg(m_kernelBlend, 3, sizeof(cl_mem), &bgImageMem[1]))
|| (CL_SUCCESS != clSetKernelArg(m_kernelBlend, 4, sizeof(cl_mem), &srcImagePtr->plane(0)))
|| (CL_SUCCESS != clSetKernelArg(m_kernelBlend, 5, sizeof(uint32_t), &crop.x))
|| (CL_SUCCESS != clSetKernelArg(m_kernelBlend, 6, sizeof(uint32_t), &crop.y))
|| (CL_SUCCESS != clSetKernelArg(m_kernelBlend, 7, sizeof(uint32_t), &crop.width))
|| (CL_SUCCESS != clSetKernelArg(m_kernelBlend, 8, sizeof(uint32_t), &crop.height))) {
ERROR("clSetKernelArg failed");
return YAMI_FAIL;
}
size_t globalWorkSize[2], localWorkSize[2];
// each work group has 8x8 work items; each work item handles 2x2 pixels
localWorkSize[0] = 8;
localWorkSize[1] = 8;
globalWorkSize[0] = ALIGN16(dst.get()->crop.width) / pixelSize;
globalWorkSize[1] = ALIGN16(dst.get()->crop.height) / 2;
if (!checkOclStatus(clEnqueueNDRangeKernel(m_context->m_queue, m_kernelBlend, 2, NULL,
globalWorkSize, localWorkSize, 0, NULL, NULL),
"EnqueueNDRangeKernel")) {
return YAMI_FAIL;
}
return ret;
}
YamiStatus
OclPostProcessBlender::process(const SharedPtr<VideoFrame>& src,
const SharedPtr<VideoFrame>& dest)
{
YamiStatus status = ensureContext("blend");
if (status != YAMI_SUCCESS)
return status;
if (src->fourcc != YAMI_FOURCC_RGBA || dest->fourcc != YAMI_FOURCC_NV12) {
ERROR("only support RGBA blending on NV12 video frame");
return YAMI_INVALID_PARAM;
}
return blend(src, dest);
}
bool OclPostProcessBlender::prepareKernels()
{
m_kernelBlend = prepareKernel("blend");
return m_kernelBlend != NULL;
}
}