obs-StreamFX/source/nvidia/cuda/nvidia-cuda-context.cpp

/*
 * Modern effects for a modern Streamer
 * Copyright (C) 2020 Michael Fabian Dirks
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA
 */

#include "nvidia-cuda-context.hpp"
#include <cassert>
#include <stdexcept>
#include "util/util-logging.hpp"

#ifdef _DEBUG
#define ST_PREFIX "<%s> "
#define D_LOG_ERROR(x, ...) P_LOG_ERROR(ST_PREFIX##x, __FUNCTION_SIG__, __VA_ARGS__)
#define D_LOG_WARNING(x, ...) P_LOG_WARN(ST_PREFIX##x, __FUNCTION_SIG__, __VA_ARGS__)
#define D_LOG_INFO(x, ...) P_LOG_INFO(ST_PREFIX##x, __FUNCTION_SIG__, __VA_ARGS__)
#define D_LOG_DEBUG(x, ...) P_LOG_DEBUG(ST_PREFIX##x, __FUNCTION_SIG__, __VA_ARGS__)
#else
#define ST_PREFIX "<nvidia::cuda::context> "
#define D_LOG_ERROR(...) P_LOG_ERROR(ST_PREFIX __VA_ARGS__)
#define D_LOG_WARNING(...) P_LOG_WARN(ST_PREFIX __VA_ARGS__)
#define D_LOG_INFO(...) P_LOG_INFO(ST_PREFIX __VA_ARGS__)
#define D_LOG_DEBUG(...) P_LOG_DEBUG(ST_PREFIX __VA_ARGS__)
#endif

#ifdef WIN32
#ifdef _MSC_VER
#pragma warning(push)
#pragma warning(disable : 4191 4242 4244 4365 4777 4986 5039 5204)
#endif
#include <atlutil.h>
#ifdef _MSC_VER
#pragma warning(pop)
#endif
#endif

#define ENABLE_STACK_CHECKS

streamfx::nvidia::cuda::context::~context()
{
	D_LOG_DEBUG("Finalizing... (Addr: 0x%" PRIuPTR ")", this);

	if (_has_device) {
		_cuda->cuDevicePrimaryCtxRelease(_device);
	} else {
		_cuda->cuCtxDestroy(_ctx);
	}
}

streamfx::nvidia::cuda::context::context()
	: _cuda(::streamfx::nvidia::cuda::cuda::get()), _ctx(), _has_device(false), _device()
{
	D_LOG_DEBUG("Initializating... (Addr: 0x%" PRIuPTR ")", this);
}

#ifdef WIN32
streamfx::nvidia::cuda::context::context(ID3D11Device* device) : context()
{
	using namespace streamfx::nvidia::cuda;

	if (!device)
		throw std::invalid_argument("device");
	// Get DXGI Device
	IDXGIDevice* dxgi_device; // Don't use ATL::CComPtr
	device->QueryInterface(__uuidof(IDXGIDevice), (void**)&dxgi_device);

	// Get DXGI Adapter
	ATL::CComPtr<IDXGIAdapter> dxgi_adapter;
	dxgi_device->GetAdapter(&dxgi_adapter);

	// Get Device Index
	if (result res = _cuda->cuD3D11GetDevice(&_device, dxgi_adapter); res != result::SUCCESS) {
		throw std::runtime_error("Failed to get device index for device.");
	}

	_cuda->cuDevicePrimaryCtxSetFlags(_device, context_flags::SCHEDULER_BLOCKING_SYNC);

	// Acquire Context
	if (result res = _cuda->cuDevicePrimaryCtxRetain(&_ctx, _device); res != result::SUCCESS) {
		throw std::runtime_error("Failed to acquire primary device context.");
	}

	// Log some information.
	std::string device_name;
	uuid_t      device_uuid;
	luid_t      device_luid;
	uint32_t    device_luid_mask;
	{
		// Device Name
		std::vector<char> name(256, 0);
		_cuda->cuDeviceGetName(name.data(), name.size() - 1, _device);
		device_name = std::string(name.data(), name.data() + strlen(name.data()));

		// Device LUID
		_cuda->cuDeviceGetLuid(&device_luid, &device_luid_mask, _device);

		// Device UUID
		_cuda->cuDeviceGetUuid(&device_uuid, _device);
	}

	D_LOG_INFO("Initialized CUDA on device '%s' (%08" PRIx32 "-%04" PRIx16 "-%04" PRIx16 "-%04" PRIx16 "-%04" PRIx16
			   "%08" PRIx32 ", %08" PRIx64 ", %" PRIu32 ").",
			   device_name.c_str(), device_uuid.uuid.a, device_uuid.uuid.b, device_uuid.uuid.c, device_uuid.uuid.d,
			   device_uuid.uuid.e, device_uuid.uuid.f, device_luid.luid, device_luid_mask);

	_has_device = true;
}
#endif

::streamfx::nvidia::cuda::context_t streamfx::nvidia::cuda::context::get()
{
	return _ctx;
}

std::shared_ptr<::streamfx::nvidia::cuda::context_stack> streamfx::nvidia::cuda::context::enter()
{
	return std::make_shared<::streamfx::nvidia::cuda::context_stack>(shared_from_this());
}

void streamfx::nvidia::cuda::context::push()
{
	if (auto res = _cuda->cuCtxPushCurrent(_ctx); res != ::streamfx::nvidia::cuda::result::SUCCESS) {
		throw ::streamfx::nvidia::cuda::cuda_error(res);
	}
}

void streamfx::nvidia::cuda::context::pop()
{
#ifdef ENABLE_STACK_CHECKS
	::streamfx::nvidia::cuda::context_t ctx;
	if (_cuda->cuCtxGetCurrent(&ctx) == ::streamfx::nvidia::cuda::result::SUCCESS)
		assert(ctx == _ctx);
#endif

	assert(_cuda->cuCtxPopCurrent(&ctx) == ::streamfx::nvidia::cuda::result::SUCCESS);
}

void streamfx::nvidia::cuda::context::synchronize()
{
	//D_LOG_DEBUG("Synchronizing... (Addr: 0x%" PRIuPTR ")", this);

#ifdef ENABLE_STACK_CHECKS
	::streamfx::nvidia::cuda::context_t ctx;
	if (_cuda->cuCtxGetCurrent(&ctx) == ::streamfx::nvidia::cuda::result::SUCCESS)
		assert(ctx == _ctx);
#endif

	if (auto res = _cuda->cuCtxSynchronize(); res != ::streamfx::nvidia::cuda::result::SUCCESS) {
		throw ::streamfx::nvidia::cuda::cuda_error(res);
	}
}