This commit is contained in:
mofeng-git
2025-12-28 18:19:16 +08:00
commit d143d158e4
771 changed files with 220548 additions and 0 deletions

View File

@@ -0,0 +1,602 @@
/*
* This copyright notice applies to this header file only:
*
* Copyright (c) 2016
*
* 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.
*/
#if !defined(FFNV_DYNLINK_CUDA_H) && !defined(CUDA_VERSION)
#define FFNV_DYNLINK_CUDA_H
#include <stddef.h>
#include <stdint.h>
#define CUDA_VERSION 7050
#if defined(_WIN32) || defined(__CYGWIN__)
#define CUDAAPI __stdcall
#else
#define CUDAAPI
#endif
#define CU_CTX_SCHED_BLOCKING_SYNC 4
typedef int CUdevice;
#if defined(__x86_64) || defined(AMD64) || defined(_M_AMD64) || defined(__LP64__) || defined(__aarch64__)
typedef unsigned long long CUdeviceptr;
#else
typedef unsigned int CUdeviceptr;
#endif
typedef unsigned long long CUtexObject;
typedef struct CUarray_st *CUarray;
typedef struct CUctx_st *CUcontext;
typedef struct CUstream_st *CUstream;
typedef struct CUevent_st *CUevent;
typedef struct CUfunc_st *CUfunction;
typedef struct CUmod_st *CUmodule;
typedef struct CUmipmappedArray_st *CUmipmappedArray;
typedef struct CUgraphicsResource_st *CUgraphicsResource;
typedef struct CUextMemory_st *CUexternalMemory;
typedef struct CUextSemaphore_st *CUexternalSemaphore;
typedef struct CUeglStreamConnection_st *CUeglStreamConnection;
typedef struct CUlinkState_st *CUlinkState;
typedef enum cudaError_enum {
CUDA_SUCCESS = 0,
CUDA_ERROR_INVALID_VALUE = 1,
CUDA_ERROR_OUT_OF_MEMORY = 2,
CUDA_ERROR_NOT_INITIALIZED = 3,
CUDA_ERROR_DEINITIALIZED = 4,
CUDA_ERROR_PROFILER_DISABLED = 5,
CUDA_ERROR_PROFILER_NOT_INITIALIZED = 6,
CUDA_ERROR_PROFILER_ALREADY_STARTED = 7,
CUDA_ERROR_PROFILER_ALREADY_STOPPED = 8,
CUDA_ERROR_STUB_LIBRARY = 34,
CUDA_ERROR_NO_DEVICE = 100,
CUDA_ERROR_INVALID_DEVICE = 101,
CUDA_ERROR_DEVICE_NOT_LICENSED = 102,
CUDA_ERROR_INVALID_IMAGE = 200,
CUDA_ERROR_INVALID_CONTEXT = 201,
CUDA_ERROR_CONTEXT_ALREADY_CURRENT = 202,
CUDA_ERROR_MAP_FAILED = 205,
CUDA_ERROR_UNMAP_FAILED = 206,
CUDA_ERROR_ARRAY_IS_MAPPED = 207,
CUDA_ERROR_ALREADY_MAPPED = 208,
CUDA_ERROR_NO_BINARY_FOR_GPU = 209,
CUDA_ERROR_ALREADY_ACQUIRED = 210,
CUDA_ERROR_NOT_MAPPED = 211,
CUDA_ERROR_NOT_MAPPED_AS_ARRAY = 212,
CUDA_ERROR_NOT_MAPPED_AS_POINTER = 213,
CUDA_ERROR_ECC_UNCORRECTABLE = 214,
CUDA_ERROR_UNSUPPORTED_LIMIT = 215,
CUDA_ERROR_CONTEXT_ALREADY_IN_USE = 216,
CUDA_ERROR_PEER_ACCESS_UNSUPPORTED = 217,
CUDA_ERROR_INVALID_PTX = 218,
CUDA_ERROR_INVALID_GRAPHICS_CONTEXT = 219,
CUDA_ERROR_NVLINK_UNCORRECTABLE = 220,
CUDA_ERROR_JIT_COMPILER_NOT_FOUND = 221,
CUDA_ERROR_UNSUPPORTED_PTX_VERSION = 222,
CUDA_ERROR_JIT_COMPILATION_DISABLED = 223,
CUDA_ERROR_UNSUPPORTED_EXEC_AFFINITY = 224,
CUDA_ERROR_INVALID_SOURCE = 300,
CUDA_ERROR_FILE_NOT_FOUND = 301,
CUDA_ERROR_SHARED_OBJECT_SYMBOL_NOT_FOUND = 302,
CUDA_ERROR_SHARED_OBJECT_INIT_FAILED = 303,
CUDA_ERROR_OPERATING_SYSTEM = 304,
CUDA_ERROR_INVALID_HANDLE = 400,
CUDA_ERROR_ILLEGAL_STATE = 401,
CUDA_ERROR_NOT_FOUND = 500,
CUDA_ERROR_NOT_READY = 600,
CUDA_ERROR_ILLEGAL_ADDRESS = 700,
CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES = 701,
CUDA_ERROR_LAUNCH_TIMEOUT = 702,
CUDA_ERROR_LAUNCH_INCOMPATIBLE_TEXTURING = 703,
CUDA_ERROR_PEER_ACCESS_ALREADY_ENABLED = 704,
CUDA_ERROR_PEER_ACCESS_NOT_ENABLED = 705,
CUDA_ERROR_PRIMARY_CONTEXT_ACTIVE = 708,
CUDA_ERROR_CONTEXT_IS_DESTROYED = 709,
CUDA_ERROR_ASSERT = 710,
CUDA_ERROR_TOO_MANY_PEERS = 711,
CUDA_ERROR_HOST_MEMORY_ALREADY_REGISTERED = 712,
CUDA_ERROR_HOST_MEMORY_NOT_REGISTERED = 713,
CUDA_ERROR_HARDWARE_STACK_ERROR = 714,
CUDA_ERROR_ILLEGAL_INSTRUCTION = 715,
CUDA_ERROR_MISALIGNED_ADDRESS = 716,
CUDA_ERROR_INVALID_ADDRESS_SPACE = 717,
CUDA_ERROR_INVALID_PC = 718,
CUDA_ERROR_LAUNCH_FAILED = 719,
CUDA_ERROR_COOPERATIVE_LAUNCH_TOO_LARGE = 720,
CUDA_ERROR_NOT_PERMITTED = 800,
CUDA_ERROR_NOT_SUPPORTED = 801,
CUDA_ERROR_SYSTEM_NOT_READY = 802,
CUDA_ERROR_SYSTEM_DRIVER_MISMATCH = 803,
CUDA_ERROR_COMPAT_NOT_SUPPORTED_ON_DEVICE = 804,
CUDA_ERROR_MPS_CONNECTION_FAILED = 805,
CUDA_ERROR_MPS_RPC_FAILURE = 806,
CUDA_ERROR_MPS_SERVER_NOT_READY = 807,
CUDA_ERROR_MPS_MAX_CLIENTS_REACHED = 808,
CUDA_ERROR_MPS_MAX_CONNECTIONS_REACHED = 809,
CUDA_ERROR_STREAM_CAPTURE_UNSUPPORTED = 900,
CUDA_ERROR_STREAM_CAPTURE_INVALIDATED = 901,
CUDA_ERROR_STREAM_CAPTURE_MERGE = 902,
CUDA_ERROR_STREAM_CAPTURE_UNMATCHED = 903,
CUDA_ERROR_STREAM_CAPTURE_UNJOINED = 904,
CUDA_ERROR_STREAM_CAPTURE_ISOLATION = 905,
CUDA_ERROR_STREAM_CAPTURE_IMPLICIT = 906,
CUDA_ERROR_CAPTURED_EVENT = 907,
CUDA_ERROR_STREAM_CAPTURE_WRONG_THREAD = 908,
CUDA_ERROR_TIMEOUT = 909,
CUDA_ERROR_GRAPH_EXEC_UPDATE_FAILURE = 910,
CUDA_ERROR_EXTERNAL_DEVICE = 911,
CUDA_ERROR_UNKNOWN = 999
} CUresult;
/**
* Device properties (subset)
*/
typedef enum CUdevice_attribute_enum {
CU_DEVICE_ATTRIBUTE_CLOCK_RATE = 13,
CU_DEVICE_ATTRIBUTE_TEXTURE_ALIGNMENT = 14,
CU_DEVICE_ATTRIBUTE_MULTIPROCESSOR_COUNT = 16,
CU_DEVICE_ATTRIBUTE_INTEGRATED = 18,
CU_DEVICE_ATTRIBUTE_CAN_MAP_HOST_MEMORY = 19,
CU_DEVICE_ATTRIBUTE_COMPUTE_MODE = 20,
CU_DEVICE_ATTRIBUTE_CONCURRENT_KERNELS = 31,
CU_DEVICE_ATTRIBUTE_PCI_BUS_ID = 33,
CU_DEVICE_ATTRIBUTE_PCI_DEVICE_ID = 34,
CU_DEVICE_ATTRIBUTE_TCC_DRIVER = 35,
CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE = 36,
CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH = 37,
CU_DEVICE_ATTRIBUTE_ASYNC_ENGINE_COUNT = 40,
CU_DEVICE_ATTRIBUTE_UNIFIED_ADDRESSING = 41,
CU_DEVICE_ATTRIBUTE_PCI_DOMAIN_ID = 50,
CU_DEVICE_ATTRIBUTE_TEXTURE_PITCH_ALIGNMENT = 51,
CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR = 75,
CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR = 76,
CU_DEVICE_ATTRIBUTE_MANAGED_MEMORY = 83,
CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD = 84,
CU_DEVICE_ATTRIBUTE_MULTI_GPU_BOARD_GROUP_ID = 85,
} CUdevice_attribute;
typedef enum CUarray_format_enum {
CU_AD_FORMAT_UNSIGNED_INT8 = 0x01,
CU_AD_FORMAT_UNSIGNED_INT16 = 0x02,
CU_AD_FORMAT_UNSIGNED_INT32 = 0x03,
CU_AD_FORMAT_SIGNED_INT8 = 0x08,
CU_AD_FORMAT_SIGNED_INT16 = 0x09,
CU_AD_FORMAT_SIGNED_INT32 = 0x0a,
CU_AD_FORMAT_HALF = 0x10,
CU_AD_FORMAT_FLOAT = 0x20
} CUarray_format;
typedef enum CUmemorytype_enum {
CU_MEMORYTYPE_HOST = 1,
CU_MEMORYTYPE_DEVICE = 2,
CU_MEMORYTYPE_ARRAY = 3
} CUmemorytype;
typedef enum CUlimit_enum {
CU_LIMIT_STACK_SIZE = 0,
CU_LIMIT_PRINTF_FIFO_SIZE = 1,
CU_LIMIT_MALLOC_HEAP_SIZE = 2,
CU_LIMIT_DEV_RUNTIME_SYNC_DEPTH = 3,
CU_LIMIT_DEV_RUNTIME_PENDING_LAUNCH_COUNT = 4
} CUlimit;
typedef enum CUresourcetype_enum {
CU_RESOURCE_TYPE_ARRAY = 0x00,
CU_RESOURCE_TYPE_MIPMAPPED_ARRAY = 0x01,
CU_RESOURCE_TYPE_LINEAR = 0x02,
CU_RESOURCE_TYPE_PITCH2D = 0x03
} CUresourcetype;
typedef enum CUaddress_mode_enum {
CU_TR_ADDRESS_MODE_WRAP = 0,
CU_TR_ADDRESS_MODE_CLAMP = 1,
CU_TR_ADDRESS_MODE_MIRROR = 2,
CU_TR_ADDRESS_MODE_BORDER = 3
} CUaddress_mode;
typedef enum CUfilter_mode_enum {
CU_TR_FILTER_MODE_POINT = 0,
CU_TR_FILTER_MODE_LINEAR = 1
} CUfilter_mode;
typedef enum CUgraphicsRegisterFlags_enum {
CU_GRAPHICS_REGISTER_FLAGS_NONE = 0,
CU_GRAPHICS_REGISTER_FLAGS_READ_ONLY = 1,
CU_GRAPHICS_REGISTER_FLAGS_WRITE_DISCARD = 2,
CU_GRAPHICS_REGISTER_FLAGS_SURFACE_LDST = 4,
CU_GRAPHICS_REGISTER_FLAGS_TEXTURE_GATHER = 8
} CUgraphicsRegisterFlags;
typedef enum CUexternalMemoryHandleType_enum {
CU_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_FD = 1,
CU_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32 = 2,
CU_EXTERNAL_MEMORY_HANDLE_TYPE_OPAQUE_WIN32_KMT = 3,
CU_EXTERNAL_MEMORY_HANDLE_TYPE_D3D12_HEAP = 4,
CU_EXTERNAL_MEMORY_HANDLE_TYPE_D3D12_RESOURCE = 5,
} CUexternalMemoryHandleType;
typedef enum CUexternalSemaphoreHandleType_enum {
CU_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_FD = 1,
CU_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32 = 2,
CU_EXTERNAL_SEMAPHORE_HANDLE_TYPE_OPAQUE_WIN32_KMT = 3,
CU_EXTERNAL_SEMAPHORE_HANDLE_TYPE_D3D12_FENCE = 4,
CU_EXTERNAL_SEMAPHORE_HANDLE_TYPE_TIMELINE_SEMAPHORE_FD = 9,
CU_EXTERNAL_SEMAPHORE_HANDLE_TYPE_TIMELINE_SEMAPHORE_WIN32 = 10,
} CUexternalSemaphoreHandleType;
typedef enum CUjit_option_enum
{
CU_JIT_MAX_REGISTERS = 0,
CU_JIT_THREADS_PER_BLOCK = 1,
CU_JIT_WALL_TIME = 2,
CU_JIT_INFO_LOG_BUFFER = 3,
CU_JIT_INFO_LOG_BUFFER_SIZE_BYTES = 4,
CU_JIT_ERROR_LOG_BUFFER = 5,
CU_JIT_ERROR_LOG_BUFFER_SIZE_BYTES = 6,
CU_JIT_OPTIMIZATION_LEVEL = 7,
CU_JIT_TARGET_FROM_CUCONTEXT = 8,
CU_JIT_TARGET = 9,
CU_JIT_FALLBACK_STRATEGY = 10,
CU_JIT_GENERATE_DEBUG_INFO = 11,
CU_JIT_LOG_VERBOSE = 12,
CU_JIT_GENERATE_LINE_INFO = 13,
CU_JIT_CACHE_MODE = 14,
CU_JIT_NEW_SM3X_OPT = 15,
CU_JIT_FAST_COMPILE = 16,
CU_JIT_GLOBAL_SYMBOL_NAMES = 17,
CU_JIT_GLOBAL_SYMBOL_ADDRESSES = 18,
CU_JIT_GLOBAL_SYMBOL_COUNT = 19,
CU_JIT_NUM_OPTIONS
} CUjit_option;
typedef enum CUjitInputType_enum
{
CU_JIT_INPUT_CUBIN = 0,
CU_JIT_INPUT_PTX = 1,
CU_JIT_INPUT_FATBINARY = 2,
CU_JIT_INPUT_OBJECT = 3,
CU_JIT_INPUT_LIBRARY = 4,
CU_JIT_NUM_INPUT_TYPES
} CUjitInputType;
typedef enum CUeglFrameType
{
CU_EGL_FRAME_TYPE_ARRAY = 0,
CU_EGL_FRAME_TYPE_PITCH = 1,
} CUeglFrameType;
typedef enum CUeglColorFormat
{
CU_EGL_COLOR_FORMAT_YUV420_PLANAR = 0x00,
CU_EGL_COLOR_FORMAT_YUV420_SEMIPLANAR = 0x01,
CU_EGL_COLOR_FORMAT_YVU420_SEMIPLANAR = 0x15,
CU_EGL_COLOR_FORMAT_Y10V10U10_420_SEMIPLANAR = 0x17,
CU_EGL_COLOR_FORMAT_Y12V12U12_420_SEMIPLANAR = 0x19,
} CUeglColorFormat;
typedef enum CUd3d11DeviceList_enum
{
CU_D3D11_DEVICE_LIST_ALL = 1,
CU_D3D11_DEVICE_LIST_CURRENT_FRAME = 2,
CU_D3D11_DEVICE_LIST_NEXT_FRAME = 3,
} CUd3d11DeviceList;
#ifndef CU_UUID_HAS_BEEN_DEFINED
#define CU_UUID_HAS_BEEN_DEFINED
typedef struct CUuuid_st {
char bytes[16];
} CUuuid;
#endif
typedef struct CUDA_MEMCPY2D_st {
size_t srcXInBytes;
size_t srcY;
CUmemorytype srcMemoryType;
const void *srcHost;
CUdeviceptr srcDevice;
CUarray srcArray;
size_t srcPitch;
size_t dstXInBytes;
size_t dstY;
CUmemorytype dstMemoryType;
void *dstHost;
CUdeviceptr dstDevice;
CUarray dstArray;
size_t dstPitch;
size_t WidthInBytes;
size_t Height;
} CUDA_MEMCPY2D;
typedef struct CUDA_RESOURCE_DESC_st {
CUresourcetype resType;
union {
struct {
CUarray hArray;
} array;
struct {
CUmipmappedArray hMipmappedArray;
} mipmap;
struct {
CUdeviceptr devPtr;
CUarray_format format;
unsigned int numChannels;
size_t sizeInBytes;
} linear;
struct {
CUdeviceptr devPtr;
CUarray_format format;
unsigned int numChannels;
size_t width;
size_t height;
size_t pitchInBytes;
} pitch2D;
struct {
int reserved[32];
} reserved;
} res;
unsigned int flags;
} CUDA_RESOURCE_DESC;
typedef struct CUDA_TEXTURE_DESC_st {
CUaddress_mode addressMode[3];
CUfilter_mode filterMode;
unsigned int flags;
unsigned int maxAnisotropy;
CUfilter_mode mipmapFilterMode;
float mipmapLevelBias;
float minMipmapLevelClamp;
float maxMipmapLevelClamp;
float borderColor[4];
int reserved[12];
} CUDA_TEXTURE_DESC;
/* Unused type */
typedef struct CUDA_RESOURCE_VIEW_DESC_st CUDA_RESOURCE_VIEW_DESC;
typedef unsigned int GLenum;
typedef unsigned int GLuint;
/*
* Prefix type name to avoid collisions. Clients using these types
* will include the real headers with real definitions.
*/
typedef int32_t ffnv_EGLint;
typedef void *ffnv_EGLStreamKHR;
typedef enum CUGLDeviceList_enum {
CU_GL_DEVICE_LIST_ALL = 1,
CU_GL_DEVICE_LIST_CURRENT_FRAME = 2,
CU_GL_DEVICE_LIST_NEXT_FRAME = 3,
} CUGLDeviceList;
typedef struct CUDA_EXTERNAL_MEMORY_HANDLE_DESC_st {
CUexternalMemoryHandleType type;
union {
int fd;
struct {
void *handle;
const void *name;
} win32;
} handle;
unsigned long long size;
unsigned int flags;
unsigned int reserved[16];
} CUDA_EXTERNAL_MEMORY_HANDLE_DESC;
typedef struct CUDA_EXTERNAL_MEMORY_BUFFER_DESC_st {
unsigned long long offset;
unsigned long long size;
unsigned int flags;
unsigned int reserved[16];
} CUDA_EXTERNAL_MEMORY_BUFFER_DESC;
typedef struct CUDA_EXTERNAL_SEMAPHORE_HANDLE_DESC_st {
CUexternalSemaphoreHandleType type;
union {
int fd;
struct {
void *handle;
const void *name;
} win32;
} handle;
unsigned int flags;
unsigned int reserved[16];
} CUDA_EXTERNAL_SEMAPHORE_HANDLE_DESC;
typedef struct CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS_st {
struct {
struct {
unsigned long long value;
} fence;
unsigned int reserved[16];
} params;
unsigned int flags;
unsigned int reserved[16];
} CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS;
typedef CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS CUDA_EXTERNAL_SEMAPHORE_WAIT_PARAMS;
typedef struct CUDA_ARRAY_DESCRIPTOR_st {
size_t Width;
size_t Height;
CUarray_format Format;
unsigned int NumChannels;
} CUDA_ARRAY_DESCRIPTOR;
typedef struct CUDA_ARRAY3D_DESCRIPTOR_st {
size_t Width;
size_t Height;
size_t Depth;
CUarray_format Format;
unsigned int NumChannels;
unsigned int Flags;
} CUDA_ARRAY3D_DESCRIPTOR;
typedef struct CUDA_EXTERNAL_MEMORY_MIPMAPPED_ARRAY_DESC_st {
unsigned long long offset;
CUDA_ARRAY3D_DESCRIPTOR arrayDesc;
unsigned int numLevels;
unsigned int reserved[16];
} CUDA_EXTERNAL_MEMORY_MIPMAPPED_ARRAY_DESC;
#define CU_EGL_FRAME_MAX_PLANES 3
typedef struct CUeglFrame_st {
union {
CUarray pArray[CU_EGL_FRAME_MAX_PLANES];
void* pPitch[CU_EGL_FRAME_MAX_PLANES];
} frame;
unsigned int width;
unsigned int height;
unsigned int depth;
unsigned int pitch;
unsigned int planeCount;
unsigned int numChannels;
CUeglFrameType frameType;
CUeglColorFormat eglColorFormat;
CUarray_format cuFormat;
} CUeglFrame;
#define CU_STREAM_DEFAULT 0
#define CU_STREAM_NON_BLOCKING 1
#define CU_EVENT_DEFAULT 0
#define CU_EVENT_BLOCKING_SYNC 1
#define CU_EVENT_DISABLE_TIMING 2
#define CU_EVENT_WAIT_DEFAULT 0
#define CU_EVENT_WAIT_EXTERNAL 1
#define CU_TRSF_READ_AS_INTEGER 1
typedef void CUDAAPI CUstreamCallback(CUstream hStream, CUresult status, void *userdata);
typedef CUresult CUDAAPI tcuInit(unsigned int Flags);
typedef CUresult CUDAAPI tcuDriverGetVersion(int *driverVersion);
typedef CUresult CUDAAPI tcuDeviceGetCount(int *count);
typedef CUresult CUDAAPI tcuDeviceGet(CUdevice *device, int ordinal);
typedef CUresult CUDAAPI tcuDeviceGetAttribute(int *pi, CUdevice_attribute attrib, CUdevice dev);
typedef CUresult CUDAAPI tcuDeviceGetName(char *name, int len, CUdevice dev);
typedef CUresult CUDAAPI tcuDeviceGetUuid(CUuuid *uuid, CUdevice dev);
typedef CUresult CUDAAPI tcuDeviceGetUuid_v2(CUuuid *uuid, CUdevice dev);
typedef CUresult CUDAAPI tcuDeviceGetLuid(char* luid, unsigned int* deviceNodeMask, CUdevice dev);
typedef CUresult CUDAAPI tcuDeviceGetByPCIBusId(CUdevice* dev, const char* pciBusId);
typedef CUresult CUDAAPI tcuDeviceGetPCIBusId(char* pciBusId, int len, CUdevice dev);
typedef CUresult CUDAAPI tcuDeviceComputeCapability(int *major, int *minor, CUdevice dev);
typedef CUresult CUDAAPI tcuCtxCreate_v2(CUcontext *pctx, unsigned int flags, CUdevice dev);
typedef CUresult CUDAAPI tcuCtxGetCurrent(CUcontext *pctx);
typedef CUresult CUDAAPI tcuCtxSetLimit(CUlimit limit, size_t value);
typedef CUresult CUDAAPI tcuCtxPushCurrent_v2(CUcontext pctx);
typedef CUresult CUDAAPI tcuCtxPopCurrent_v2(CUcontext *pctx);
typedef CUresult CUDAAPI tcuCtxDestroy_v2(CUcontext ctx);
typedef CUresult CUDAAPI tcuMemAlloc_v2(CUdeviceptr *dptr, size_t bytesize);
typedef CUresult CUDAAPI tcuMemAllocPitch_v2(CUdeviceptr *dptr, size_t *pPitch, size_t WidthInBytes, size_t Height, unsigned int ElementSizeBytes);
typedef CUresult CUDAAPI tcuMemAllocManaged(CUdeviceptr *dptr, size_t bytesize, unsigned int flags);
typedef CUresult CUDAAPI tcuMemsetD8Async(CUdeviceptr dstDevice, unsigned char uc, size_t N, CUstream hStream);
typedef CUresult CUDAAPI tcuMemFree_v2(CUdeviceptr dptr);
typedef CUresult CUDAAPI tcuMemcpy(CUdeviceptr dst, CUdeviceptr src, size_t bytesize);
typedef CUresult CUDAAPI tcuMemcpyAsync(CUdeviceptr dst, CUdeviceptr src, size_t bytesize, CUstream hStream);
typedef CUresult CUDAAPI tcuMemcpy2D_v2(const CUDA_MEMCPY2D *pcopy);
typedef CUresult CUDAAPI tcuMemcpy2DAsync_v2(const CUDA_MEMCPY2D *pcopy, CUstream hStream);
typedef CUresult CUDAAPI tcuMemcpyHtoD_v2(CUdeviceptr dstDevice, const void *srcHost, size_t ByteCount);
typedef CUresult CUDAAPI tcuMemcpyHtoDAsync_v2(CUdeviceptr dstDevice, const void *srcHost, size_t ByteCount, CUstream hStream);
typedef CUresult CUDAAPI tcuMemcpyDtoH_v2(void *dstHost, CUdeviceptr srcDevice, size_t ByteCount);
typedef CUresult CUDAAPI tcuMemcpyDtoHAsync_v2(void *dstHost, CUdeviceptr srcDevice, size_t ByteCount, CUstream hStream);
typedef CUresult CUDAAPI tcuMemcpyDtoD_v2(CUdeviceptr dstDevice, CUdeviceptr srcDevice, size_t ByteCount);
typedef CUresult CUDAAPI tcuMemcpyDtoDAsync_v2(CUdeviceptr dstDevice, CUdeviceptr srcDevice, size_t ByteCount, CUstream hStream);
typedef CUresult CUDAAPI tcuGetErrorName(CUresult error, const char** pstr);
typedef CUresult CUDAAPI tcuGetErrorString(CUresult error, const char** pstr);
typedef CUresult CUDAAPI tcuCtxGetDevice(CUdevice *device);
typedef CUresult CUDAAPI tcuDevicePrimaryCtxRetain(CUcontext *pctx, CUdevice dev);
typedef CUresult CUDAAPI tcuDevicePrimaryCtxRelease(CUdevice dev);
typedef CUresult CUDAAPI tcuDevicePrimaryCtxSetFlags(CUdevice dev, unsigned int flags);
typedef CUresult CUDAAPI tcuDevicePrimaryCtxGetState(CUdevice dev, unsigned int *flags, int *active);
typedef CUresult CUDAAPI tcuDevicePrimaryCtxReset(CUdevice dev);
typedef CUresult CUDAAPI tcuStreamCreate(CUstream *phStream, unsigned int flags);
typedef CUresult CUDAAPI tcuStreamQuery(CUstream hStream);
typedef CUresult CUDAAPI tcuStreamSynchronize(CUstream hStream);
typedef CUresult CUDAAPI tcuStreamDestroy_v2(CUstream hStream);
typedef CUresult CUDAAPI tcuStreamAddCallback(CUstream hStream, CUstreamCallback *callback, void *userdata, unsigned int flags);
typedef CUresult CUDAAPI tcuStreamWaitEvent(CUstream hStream, CUevent hEvent, unsigned int flags);
typedef CUresult CUDAAPI tcuEventCreate(CUevent *phEvent, unsigned int flags);
typedef CUresult CUDAAPI tcuEventDestroy_v2(CUevent hEvent);
typedef CUresult CUDAAPI tcuEventSynchronize(CUevent hEvent);
typedef CUresult CUDAAPI tcuEventQuery(CUevent hEvent);
typedef CUresult CUDAAPI tcuEventRecord(CUevent hEvent, CUstream hStream);
typedef CUresult CUDAAPI tcuLaunchKernel(CUfunction f, unsigned int gridDimX, unsigned int gridDimY, unsigned int gridDimZ, unsigned int blockDimX, unsigned int blockDimY, unsigned int blockDimZ, unsigned int sharedMemBytes, CUstream hStream, void** kernelParams, void** extra);
typedef CUresult CUDAAPI tcuLinkCreate(unsigned int numOptions, CUjit_option* options, void** optionValues, CUlinkState* stateOut);
typedef CUresult CUDAAPI tcuLinkAddData(CUlinkState state, CUjitInputType type, void* data, size_t size, const char* name, unsigned int numOptions, CUjit_option* options, void** optionValues);
typedef CUresult CUDAAPI tcuLinkComplete(CUlinkState state, void** cubinOut, size_t* sizeOut);
typedef CUresult CUDAAPI tcuLinkDestroy(CUlinkState state);
typedef CUresult CUDAAPI tcuModuleLoadData(CUmodule* module, const void* image);
typedef CUresult CUDAAPI tcuModuleUnload(CUmodule hmod);
typedef CUresult CUDAAPI tcuModuleGetFunction(CUfunction* hfunc, CUmodule hmod, const char* name);
typedef CUresult CUDAAPI tcuModuleGetGlobal(CUdeviceptr *dptr, size_t *bytes, CUmodule hmod, const char* name);
typedef CUresult CUDAAPI tcuTexObjectCreate(CUtexObject* pTexObject, const CUDA_RESOURCE_DESC* pResDesc, const CUDA_TEXTURE_DESC* pTexDesc, const CUDA_RESOURCE_VIEW_DESC* pResViewDesc);
typedef CUresult CUDAAPI tcuTexObjectDestroy(CUtexObject texObject);
typedef CUresult CUDAAPI tcuGLGetDevices_v2(unsigned int* pCudaDeviceCount, CUdevice* pCudaDevices, unsigned int cudaDeviceCount, CUGLDeviceList deviceList);
typedef CUresult CUDAAPI tcuGraphicsGLRegisterImage(CUgraphicsResource* pCudaResource, GLuint image, GLenum target, unsigned int Flags);
typedef CUresult CUDAAPI tcuGraphicsUnregisterResource(CUgraphicsResource resource);
typedef CUresult CUDAAPI tcuGraphicsMapResources(unsigned int count, CUgraphicsResource* resources, CUstream hStream);
typedef CUresult CUDAAPI tcuGraphicsUnmapResources(unsigned int count, CUgraphicsResource* resources, CUstream hStream);
typedef CUresult CUDAAPI tcuGraphicsSubResourceGetMappedArray(CUarray* pArray, CUgraphicsResource resource, unsigned int arrayIndex, unsigned int mipLevel);
typedef CUresult CUDAAPI tcuGraphicsResourceGetMappedPointer(CUdeviceptr *devPtrOut, size_t *sizeOut, CUgraphicsResource resource);
typedef CUresult CUDAAPI tcuGraphicsResourceSetMapFlags_v2(CUgraphicsResource resource, unsigned int flags);
typedef CUresult CUDAAPI tcuImportExternalMemory(CUexternalMemory* extMem_out, const CUDA_EXTERNAL_MEMORY_HANDLE_DESC* memHandleDesc);
typedef CUresult CUDAAPI tcuDestroyExternalMemory(CUexternalMemory extMem);
typedef CUresult CUDAAPI tcuExternalMemoryGetMappedBuffer(CUdeviceptr* devPtr, CUexternalMemory extMem, const CUDA_EXTERNAL_MEMORY_BUFFER_DESC* bufferDesc);
typedef CUresult CUDAAPI tcuExternalMemoryGetMappedMipmappedArray(CUmipmappedArray* mipmap, CUexternalMemory extMem, const CUDA_EXTERNAL_MEMORY_MIPMAPPED_ARRAY_DESC* mipmapDesc);
typedef CUresult CUDAAPI tcuMipmappedArrayGetLevel(CUarray* pLevelArray, CUmipmappedArray hMipmappedArray, unsigned int level);
typedef CUresult CUDAAPI tcuMipmappedArrayDestroy(CUmipmappedArray hMipmappedArray);
typedef CUresult CUDAAPI tcuImportExternalSemaphore(CUexternalSemaphore* extSem_out, const CUDA_EXTERNAL_SEMAPHORE_HANDLE_DESC* semHandleDesc);
typedef CUresult CUDAAPI tcuDestroyExternalSemaphore(CUexternalSemaphore extSem);
typedef CUresult CUDAAPI tcuSignalExternalSemaphoresAsync(const CUexternalSemaphore* extSemArray, const CUDA_EXTERNAL_SEMAPHORE_SIGNAL_PARAMS* paramsArray, unsigned int numExtSems, CUstream stream);
typedef CUresult CUDAAPI tcuWaitExternalSemaphoresAsync(const CUexternalSemaphore* extSemArray, const CUDA_EXTERNAL_SEMAPHORE_WAIT_PARAMS* paramsArray, unsigned int numExtSems, CUstream stream);
typedef CUresult CUDAAPI tcuArrayCreate(CUarray *pHandle, const CUDA_ARRAY_DESCRIPTOR* pAllocateArray);
typedef CUresult CUDAAPI tcuArray3DCreate(CUarray *pHandle, const CUDA_ARRAY3D_DESCRIPTOR* pAllocateArray);
typedef CUresult CUDAAPI tcuArrayDestroy(CUarray hArray);
typedef CUresult CUDAAPI tcuEGLStreamProducerConnect(CUeglStreamConnection* conn, ffnv_EGLStreamKHR stream, ffnv_EGLint width, ffnv_EGLint height);
typedef CUresult CUDAAPI tcuEGLStreamProducerDisconnect(CUeglStreamConnection* conn);
typedef CUresult CUDAAPI tcuEGLStreamConsumerDisconnect(CUeglStreamConnection* conn);
typedef CUresult CUDAAPI tcuEGLStreamProducerPresentFrame(CUeglStreamConnection* conn, CUeglFrame eglframe, CUstream* pStream);
typedef CUresult CUDAAPI tcuEGLStreamProducerReturnFrame(CUeglStreamConnection* conn, CUeglFrame* eglframe, CUstream* pStream);
typedef CUresult CUDAAPI tcuD3D11GetDevice(CUdevice *device, void *dxgiAdapter);
typedef CUresult CUDAAPI tcuD3D11GetDevices(unsigned int *deviceCountOut, CUdevice *devices, unsigned int deviceCount, void *d3d11device, CUd3d11DeviceList listType);
typedef CUresult CUDAAPI tcuGraphicsD3D11RegisterResource(CUgraphicsResource *cudaResourceOut, void *d3d11Resource, unsigned int flags);
#endif

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,481 @@
/*
* This copyright notice applies to this header file only:
*
* Copyright (c) 2016
*
* 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.
*/
#ifndef FFNV_CUDA_DYNLINK_LOADER_H
#define FFNV_CUDA_DYNLINK_LOADER_H
#include <stdlib.h>
#include "dynlink_cuda.h"
#include "dynlink_nvcuvid.h"
#include "nvEncodeAPI.h"
#if defined(_WIN32) && (!defined(FFNV_LOAD_FUNC) || !defined(FFNV_SYM_FUNC) || !defined(FFNV_LIB_HANDLE))
# include <windows.h>
#endif
#ifndef FFNV_LIB_HANDLE
# if defined(_WIN32)
# define FFNV_LIB_HANDLE HMODULE
# else
# define FFNV_LIB_HANDLE void*
# endif
#endif
#if defined(_WIN32) || defined(__CYGWIN__)
# define CUDA_LIBNAME "nvcuda.dll"
# define NVCUVID_LIBNAME "nvcuvid.dll"
# if defined(_WIN64) || defined(__CYGWIN64__)
# define NVENC_LIBNAME "nvEncodeAPI64.dll"
# else
# define NVENC_LIBNAME "nvEncodeAPI.dll"
# endif
#else
# define CUDA_LIBNAME "libcuda.so.1"
# define NVCUVID_LIBNAME "libnvcuvid.so.1"
# define NVENC_LIBNAME "libnvidia-encode.so.1"
#endif
#if !defined(FFNV_LOAD_FUNC) || !defined(FFNV_SYM_FUNC)
# ifdef _WIN32
# define FFNV_LOAD_FUNC(path) LoadLibrary(TEXT(path))
# define FFNV_SYM_FUNC(lib, sym) GetProcAddress((lib), (sym))
# define FFNV_FREE_FUNC(lib) FreeLibrary(lib)
# else
# include <dlfcn.h>
# define FFNV_LOAD_FUNC(path) dlopen((path), RTLD_LAZY)
# define FFNV_SYM_FUNC(lib, sym) dlsym((lib), (sym))
# define FFNV_FREE_FUNC(lib) dlclose(lib)
# endif
#endif
#if !defined(FFNV_LOG_FUNC) || !defined(FFNV_DEBUG_LOG_FUNC)
# include <stdio.h>
# define FFNV_LOG_FUNC(logctx, msg, ...) fprintf(stderr, (msg), __VA_ARGS__)
# define FFNV_DEBUG_LOG_FUNC(logctx, msg, ...)
#endif
#define LOAD_LIBRARY(l, path) \
do { \
if (!((l) = FFNV_LOAD_FUNC(path))) { \
FFNV_LOG_FUNC(logctx, "Cannot load %s\n", path); \
ret = -1; \
goto error; \
} \
FFNV_DEBUG_LOG_FUNC(logctx, "Loaded lib: %s\n", path); \
} while (0)
#define LOAD_SYMBOL(fun, tp, symbol) \
do { \
if (!((f->fun) = (tp*)FFNV_SYM_FUNC(f->lib, symbol))) { \
FFNV_LOG_FUNC(logctx, "Cannot load %s\n", symbol); \
ret = -1; \
goto error; \
} \
FFNV_DEBUG_LOG_FUNC(logctx, "Loaded sym: %s\n", symbol); \
} while (0)
#define LOAD_SYMBOL_OPT(fun, tp, symbol) \
do { \
if (!((f->fun) = (tp*)FFNV_SYM_FUNC(f->lib, symbol))) { \
FFNV_DEBUG_LOG_FUNC(logctx, "Cannot load optional %s\n", symbol); \
} else { \
FFNV_DEBUG_LOG_FUNC(logctx, "Loaded sym: %s\n", symbol); \
} \
} while (0)
#define GENERIC_LOAD_FUNC_PREAMBLE(T, n, N) \
T *f; \
int ret; \
\
n##_free_functions(functions); \
\
f = *functions = (T*)calloc(1, sizeof(*f)); \
if (!f) \
return -1; \
\
LOAD_LIBRARY(f->lib, N);
#define GENERIC_LOAD_FUNC_FINALE(n) \
return 0; \
error: \
n##_free_functions(functions); \
return ret;
#define GENERIC_FREE_FUNC() \
if (!functions) \
return; \
if (*functions && (*functions)->lib) \
FFNV_FREE_FUNC((*functions)->lib); \
free(*functions); \
*functions = NULL;
#ifdef FFNV_DYNLINK_CUDA_H
typedef struct CudaFunctions {
tcuInit *cuInit;
tcuDriverGetVersion *cuDriverGetVersion;
tcuDeviceGetCount *cuDeviceGetCount;
tcuDeviceGet *cuDeviceGet;
tcuDeviceGetAttribute *cuDeviceGetAttribute;
tcuDeviceGetName *cuDeviceGetName;
tcuDeviceGetUuid *cuDeviceGetUuid;
tcuDeviceGetUuid_v2 *cuDeviceGetUuid_v2;
tcuDeviceGetLuid *cuDeviceGetLuid;
tcuDeviceGetByPCIBusId *cuDeviceGetByPCIBusId;
tcuDeviceGetPCIBusId *cuDeviceGetPCIBusId;
tcuDeviceComputeCapability *cuDeviceComputeCapability;
tcuCtxCreate_v2 *cuCtxCreate;
tcuCtxGetCurrent *cuCtxGetCurrent;
tcuCtxSetLimit *cuCtxSetLimit;
tcuCtxPushCurrent_v2 *cuCtxPushCurrent;
tcuCtxPopCurrent_v2 *cuCtxPopCurrent;
tcuCtxDestroy_v2 *cuCtxDestroy;
tcuMemAlloc_v2 *cuMemAlloc;
tcuMemAllocPitch_v2 *cuMemAllocPitch;
tcuMemAllocManaged *cuMemAllocManaged;
tcuMemsetD8Async *cuMemsetD8Async;
tcuMemFree_v2 *cuMemFree;
tcuMemcpy *cuMemcpy;
tcuMemcpyAsync *cuMemcpyAsync;
tcuMemcpy2D_v2 *cuMemcpy2D;
tcuMemcpy2DAsync_v2 *cuMemcpy2DAsync;
tcuMemcpyHtoD_v2 *cuMemcpyHtoD;
tcuMemcpyHtoDAsync_v2 *cuMemcpyHtoDAsync;
tcuMemcpyDtoH_v2 *cuMemcpyDtoH;
tcuMemcpyDtoHAsync_v2 *cuMemcpyDtoHAsync;
tcuMemcpyDtoD_v2 *cuMemcpyDtoD;
tcuMemcpyDtoDAsync_v2 *cuMemcpyDtoDAsync;
tcuGetErrorName *cuGetErrorName;
tcuGetErrorString *cuGetErrorString;
tcuCtxGetDevice *cuCtxGetDevice;
tcuDevicePrimaryCtxRetain *cuDevicePrimaryCtxRetain;
tcuDevicePrimaryCtxRelease *cuDevicePrimaryCtxRelease;
tcuDevicePrimaryCtxSetFlags *cuDevicePrimaryCtxSetFlags;
tcuDevicePrimaryCtxGetState *cuDevicePrimaryCtxGetState;
tcuDevicePrimaryCtxReset *cuDevicePrimaryCtxReset;
tcuStreamCreate *cuStreamCreate;
tcuStreamQuery *cuStreamQuery;
tcuStreamSynchronize *cuStreamSynchronize;
tcuStreamDestroy_v2 *cuStreamDestroy;
tcuStreamAddCallback *cuStreamAddCallback;
tcuStreamWaitEvent *cuStreamWaitEvent;
tcuEventCreate *cuEventCreate;
tcuEventDestroy_v2 *cuEventDestroy;
tcuEventSynchronize *cuEventSynchronize;
tcuEventQuery *cuEventQuery;
tcuEventRecord *cuEventRecord;
tcuLaunchKernel *cuLaunchKernel;
tcuLinkCreate *cuLinkCreate;
tcuLinkAddData *cuLinkAddData;
tcuLinkComplete *cuLinkComplete;
tcuLinkDestroy *cuLinkDestroy;
tcuModuleLoadData *cuModuleLoadData;
tcuModuleUnload *cuModuleUnload;
tcuModuleGetFunction *cuModuleGetFunction;
tcuModuleGetGlobal *cuModuleGetGlobal;
tcuTexObjectCreate *cuTexObjectCreate;
tcuTexObjectDestroy *cuTexObjectDestroy;
tcuGLGetDevices_v2 *cuGLGetDevices;
tcuGraphicsGLRegisterImage *cuGraphicsGLRegisterImage;
tcuGraphicsUnregisterResource *cuGraphicsUnregisterResource;
tcuGraphicsMapResources *cuGraphicsMapResources;
tcuGraphicsUnmapResources *cuGraphicsUnmapResources;
tcuGraphicsSubResourceGetMappedArray *cuGraphicsSubResourceGetMappedArray;
tcuGraphicsResourceGetMappedPointer *cuGraphicsResourceGetMappedPointer;
tcuGraphicsResourceSetMapFlags_v2 *cuGraphicsResourceSetMapFlags;
tcuImportExternalMemory *cuImportExternalMemory;
tcuDestroyExternalMemory *cuDestroyExternalMemory;
tcuExternalMemoryGetMappedBuffer *cuExternalMemoryGetMappedBuffer;
tcuExternalMemoryGetMappedMipmappedArray *cuExternalMemoryGetMappedMipmappedArray;
tcuMipmappedArrayDestroy *cuMipmappedArrayDestroy;
tcuMipmappedArrayGetLevel *cuMipmappedArrayGetLevel;
tcuImportExternalSemaphore *cuImportExternalSemaphore;
tcuDestroyExternalSemaphore *cuDestroyExternalSemaphore;
tcuSignalExternalSemaphoresAsync *cuSignalExternalSemaphoresAsync;
tcuWaitExternalSemaphoresAsync *cuWaitExternalSemaphoresAsync;
tcuArrayCreate *cuArrayCreate;
tcuArray3DCreate *cuArray3DCreate;
tcuArrayDestroy *cuArrayDestroy;
tcuEGLStreamProducerConnect *cuEGLStreamProducerConnect;
tcuEGLStreamProducerDisconnect *cuEGLStreamProducerDisconnect;
tcuEGLStreamConsumerDisconnect *cuEGLStreamConsumerDisconnect;
tcuEGLStreamProducerPresentFrame *cuEGLStreamProducerPresentFrame;
tcuEGLStreamProducerReturnFrame *cuEGLStreamProducerReturnFrame;
#if defined(_WIN32) || defined(__CYGWIN__)
tcuD3D11GetDevice *cuD3D11GetDevice;
tcuD3D11GetDevices *cuD3D11GetDevices;
tcuGraphicsD3D11RegisterResource *cuGraphicsD3D11RegisterResource;
#endif
FFNV_LIB_HANDLE lib;
} CudaFunctions;
#else
typedef struct CudaFunctions CudaFunctions;
#endif
typedef struct CuvidFunctions {
tcuvidGetDecoderCaps *cuvidGetDecoderCaps;
tcuvidCreateDecoder *cuvidCreateDecoder;
tcuvidDestroyDecoder *cuvidDestroyDecoder;
tcuvidDecodePicture *cuvidDecodePicture;
tcuvidGetDecodeStatus *cuvidGetDecodeStatus;
tcuvidReconfigureDecoder *cuvidReconfigureDecoder;
tcuvidMapVideoFrame *cuvidMapVideoFrame;
tcuvidUnmapVideoFrame *cuvidUnmapVideoFrame;
tcuvidCtxLockCreate *cuvidCtxLockCreate;
tcuvidCtxLockDestroy *cuvidCtxLockDestroy;
tcuvidCtxLock *cuvidCtxLock;
tcuvidCtxUnlock *cuvidCtxUnlock;
#if !defined(__APPLE__)
tcuvidCreateVideoSource *cuvidCreateVideoSource;
tcuvidCreateVideoSourceW *cuvidCreateVideoSourceW;
tcuvidDestroyVideoSource *cuvidDestroyVideoSource;
tcuvidSetVideoSourceState *cuvidSetVideoSourceState;
tcuvidGetVideoSourceState *cuvidGetVideoSourceState;
tcuvidGetSourceVideoFormat *cuvidGetSourceVideoFormat;
tcuvidGetSourceAudioFormat *cuvidGetSourceAudioFormat;
#endif
tcuvidCreateVideoParser *cuvidCreateVideoParser;
tcuvidParseVideoData *cuvidParseVideoData;
tcuvidDestroyVideoParser *cuvidDestroyVideoParser;
FFNV_LIB_HANDLE lib;
} CuvidFunctions;
typedef NVENCSTATUS NVENCAPI tNvEncodeAPICreateInstance(NV_ENCODE_API_FUNCTION_LIST *functionList);
typedef NVENCSTATUS NVENCAPI tNvEncodeAPIGetMaxSupportedVersion(uint32_t* version);
typedef struct NvencFunctions {
tNvEncodeAPICreateInstance *NvEncodeAPICreateInstance;
tNvEncodeAPIGetMaxSupportedVersion *NvEncodeAPIGetMaxSupportedVersion;
FFNV_LIB_HANDLE lib;
} NvencFunctions;
#ifdef FFNV_DYNLINK_CUDA_H
static inline void cuda_free_functions(CudaFunctions **functions)
{
GENERIC_FREE_FUNC();
}
#endif
static inline void cuvid_free_functions(CuvidFunctions **functions)
{
GENERIC_FREE_FUNC();
}
static inline void nvenc_free_functions(NvencFunctions **functions)
{
GENERIC_FREE_FUNC();
}
#ifdef FFNV_DYNLINK_CUDA_H
static inline int cuda_load_functions(CudaFunctions **functions, void *logctx)
{
(void)logctx;
GENERIC_LOAD_FUNC_PREAMBLE(CudaFunctions, cuda, CUDA_LIBNAME);
LOAD_SYMBOL(cuInit, tcuInit, "cuInit");
LOAD_SYMBOL(cuDriverGetVersion, tcuDriverGetVersion, "cuDriverGetVersion");
LOAD_SYMBOL(cuDeviceGetCount, tcuDeviceGetCount, "cuDeviceGetCount");
LOAD_SYMBOL(cuDeviceGet, tcuDeviceGet, "cuDeviceGet");
LOAD_SYMBOL(cuDeviceGetAttribute, tcuDeviceGetAttribute, "cuDeviceGetAttribute");
LOAD_SYMBOL(cuDeviceGetName, tcuDeviceGetName, "cuDeviceGetName");
LOAD_SYMBOL(cuDeviceComputeCapability, tcuDeviceComputeCapability, "cuDeviceComputeCapability");
LOAD_SYMBOL(cuCtxCreate, tcuCtxCreate_v2, "cuCtxCreate_v2");
LOAD_SYMBOL(cuCtxGetCurrent, tcuCtxGetCurrent, "cuCtxGetCurrent");
LOAD_SYMBOL(cuCtxSetLimit, tcuCtxSetLimit, "cuCtxSetLimit");
LOAD_SYMBOL(cuCtxPushCurrent, tcuCtxPushCurrent_v2, "cuCtxPushCurrent_v2");
LOAD_SYMBOL(cuCtxPopCurrent, tcuCtxPopCurrent_v2, "cuCtxPopCurrent_v2");
LOAD_SYMBOL(cuCtxDestroy, tcuCtxDestroy_v2, "cuCtxDestroy_v2");
LOAD_SYMBOL(cuMemAlloc, tcuMemAlloc_v2, "cuMemAlloc_v2");
LOAD_SYMBOL(cuMemAllocPitch, tcuMemAllocPitch_v2, "cuMemAllocPitch_v2");
LOAD_SYMBOL(cuMemAllocManaged, tcuMemAllocManaged, "cuMemAllocManaged");
LOAD_SYMBOL(cuMemsetD8Async, tcuMemsetD8Async, "cuMemsetD8Async");
LOAD_SYMBOL(cuMemFree, tcuMemFree_v2, "cuMemFree_v2");
LOAD_SYMBOL(cuMemcpy, tcuMemcpy, "cuMemcpy");
LOAD_SYMBOL(cuMemcpyAsync, tcuMemcpyAsync, "cuMemcpyAsync");
LOAD_SYMBOL(cuMemcpy2D, tcuMemcpy2D_v2, "cuMemcpy2D_v2");
LOAD_SYMBOL(cuMemcpy2DAsync, tcuMemcpy2DAsync_v2, "cuMemcpy2DAsync_v2");
LOAD_SYMBOL(cuMemcpyHtoD, tcuMemcpyHtoD_v2, "cuMemcpyHtoD_v2");
LOAD_SYMBOL(cuMemcpyHtoDAsync, tcuMemcpyHtoDAsync_v2, "cuMemcpyHtoDAsync_v2");
LOAD_SYMBOL(cuMemcpyDtoH, tcuMemcpyDtoH_v2, "cuMemcpyDtoH_v2");
LOAD_SYMBOL(cuMemcpyDtoHAsync, tcuMemcpyDtoHAsync_v2, "cuMemcpyDtoHAsync_v2");
LOAD_SYMBOL(cuMemcpyDtoD, tcuMemcpyDtoD_v2, "cuMemcpyDtoD_v2");
LOAD_SYMBOL(cuMemcpyDtoDAsync, tcuMemcpyDtoDAsync_v2, "cuMemcpyDtoDAsync_v2");
LOAD_SYMBOL(cuGetErrorName, tcuGetErrorName, "cuGetErrorName");
LOAD_SYMBOL(cuGetErrorString, tcuGetErrorString, "cuGetErrorString");
LOAD_SYMBOL(cuCtxGetDevice, tcuCtxGetDevice, "cuCtxGetDevice");
LOAD_SYMBOL(cuDevicePrimaryCtxRetain, tcuDevicePrimaryCtxRetain, "cuDevicePrimaryCtxRetain");
LOAD_SYMBOL(cuDevicePrimaryCtxRelease, tcuDevicePrimaryCtxRelease, "cuDevicePrimaryCtxRelease");
LOAD_SYMBOL(cuDevicePrimaryCtxSetFlags, tcuDevicePrimaryCtxSetFlags, "cuDevicePrimaryCtxSetFlags");
LOAD_SYMBOL(cuDevicePrimaryCtxGetState, tcuDevicePrimaryCtxGetState, "cuDevicePrimaryCtxGetState");
LOAD_SYMBOL(cuDevicePrimaryCtxReset, tcuDevicePrimaryCtxReset, "cuDevicePrimaryCtxReset");
LOAD_SYMBOL(cuStreamCreate, tcuStreamCreate, "cuStreamCreate");
LOAD_SYMBOL(cuStreamQuery, tcuStreamQuery, "cuStreamQuery");
LOAD_SYMBOL(cuStreamSynchronize, tcuStreamSynchronize, "cuStreamSynchronize");
LOAD_SYMBOL(cuStreamDestroy, tcuStreamDestroy_v2, "cuStreamDestroy_v2");
LOAD_SYMBOL(cuStreamAddCallback, tcuStreamAddCallback, "cuStreamAddCallback");
LOAD_SYMBOL(cuStreamWaitEvent, tcuStreamWaitEvent, "cuStreamWaitEvent");
LOAD_SYMBOL(cuEventCreate, tcuEventCreate, "cuEventCreate");
LOAD_SYMBOL(cuEventDestroy, tcuEventDestroy_v2, "cuEventDestroy_v2");
LOAD_SYMBOL(cuEventSynchronize, tcuEventSynchronize, "cuEventSynchronize");
LOAD_SYMBOL(cuEventQuery, tcuEventQuery, "cuEventQuery");
LOAD_SYMBOL(cuEventRecord, tcuEventRecord, "cuEventRecord");
LOAD_SYMBOL(cuLaunchKernel, tcuLaunchKernel, "cuLaunchKernel");
LOAD_SYMBOL(cuLinkCreate, tcuLinkCreate, "cuLinkCreate");
LOAD_SYMBOL(cuLinkAddData, tcuLinkAddData, "cuLinkAddData");
LOAD_SYMBOL(cuLinkComplete, tcuLinkComplete, "cuLinkComplete");
LOAD_SYMBOL(cuLinkDestroy, tcuLinkDestroy, "cuLinkDestroy");
LOAD_SYMBOL(cuModuleLoadData, tcuModuleLoadData, "cuModuleLoadData");
LOAD_SYMBOL(cuModuleUnload, tcuModuleUnload, "cuModuleUnload");
LOAD_SYMBOL(cuModuleGetFunction, tcuModuleGetFunction, "cuModuleGetFunction");
LOAD_SYMBOL(cuModuleGetGlobal, tcuModuleGetGlobal, "cuModuleGetGlobal");
LOAD_SYMBOL(cuTexObjectCreate, tcuTexObjectCreate, "cuTexObjectCreate");
LOAD_SYMBOL(cuTexObjectDestroy, tcuTexObjectDestroy, "cuTexObjectDestroy");
LOAD_SYMBOL(cuGLGetDevices, tcuGLGetDevices_v2, "cuGLGetDevices_v2");
LOAD_SYMBOL(cuGraphicsGLRegisterImage, tcuGraphicsGLRegisterImage, "cuGraphicsGLRegisterImage");
LOAD_SYMBOL(cuGraphicsUnregisterResource, tcuGraphicsUnregisterResource, "cuGraphicsUnregisterResource");
LOAD_SYMBOL(cuGraphicsMapResources, tcuGraphicsMapResources, "cuGraphicsMapResources");
LOAD_SYMBOL(cuGraphicsUnmapResources, tcuGraphicsUnmapResources, "cuGraphicsUnmapResources");
LOAD_SYMBOL(cuGraphicsSubResourceGetMappedArray, tcuGraphicsSubResourceGetMappedArray, "cuGraphicsSubResourceGetMappedArray");
LOAD_SYMBOL(cuGraphicsResourceGetMappedPointer, tcuGraphicsResourceGetMappedPointer, "cuGraphicsResourceGetMappedPointer_v2");
LOAD_SYMBOL(cuGraphicsResourceSetMapFlags, tcuGraphicsResourceSetMapFlags_v2, "cuGraphicsResourceSetMapFlags_v2");
LOAD_SYMBOL_OPT(cuDeviceGetUuid, tcuDeviceGetUuid, "cuDeviceGetUuid");
LOAD_SYMBOL_OPT(cuDeviceGetUuid_v2, tcuDeviceGetUuid_v2, "cuDeviceGetUuid_v2");
LOAD_SYMBOL_OPT(cuDeviceGetLuid, tcuDeviceGetLuid, "cuDeviceGetLuid");
LOAD_SYMBOL_OPT(cuDeviceGetByPCIBusId, tcuDeviceGetByPCIBusId, "cuDeviceGetByPCIBusId");
LOAD_SYMBOL_OPT(cuDeviceGetPCIBusId, tcuDeviceGetPCIBusId, "cuDeviceGetPCIBusId");
LOAD_SYMBOL_OPT(cuImportExternalMemory, tcuImportExternalMemory, "cuImportExternalMemory");
LOAD_SYMBOL_OPT(cuDestroyExternalMemory, tcuDestroyExternalMemory, "cuDestroyExternalMemory");
LOAD_SYMBOL_OPT(cuExternalMemoryGetMappedBuffer, tcuExternalMemoryGetMappedBuffer, "cuExternalMemoryGetMappedBuffer");
LOAD_SYMBOL_OPT(cuExternalMemoryGetMappedMipmappedArray, tcuExternalMemoryGetMappedMipmappedArray, "cuExternalMemoryGetMappedMipmappedArray");
LOAD_SYMBOL_OPT(cuMipmappedArrayGetLevel, tcuMipmappedArrayGetLevel, "cuMipmappedArrayGetLevel");
LOAD_SYMBOL_OPT(cuMipmappedArrayDestroy, tcuMipmappedArrayDestroy, "cuMipmappedArrayDestroy");
LOAD_SYMBOL_OPT(cuImportExternalSemaphore, tcuImportExternalSemaphore, "cuImportExternalSemaphore");
LOAD_SYMBOL_OPT(cuDestroyExternalSemaphore, tcuDestroyExternalSemaphore, "cuDestroyExternalSemaphore");
LOAD_SYMBOL_OPT(cuSignalExternalSemaphoresAsync, tcuSignalExternalSemaphoresAsync, "cuSignalExternalSemaphoresAsync");
LOAD_SYMBOL_OPT(cuWaitExternalSemaphoresAsync, tcuWaitExternalSemaphoresAsync, "cuWaitExternalSemaphoresAsync");
LOAD_SYMBOL(cuArrayCreate, tcuArrayCreate, "cuArrayCreate_v2");
LOAD_SYMBOL(cuArray3DCreate, tcuArray3DCreate, "cuArray3DCreate_v2");
LOAD_SYMBOL(cuArrayDestroy, tcuArrayDestroy, "cuArrayDestroy");
LOAD_SYMBOL_OPT(cuEGLStreamProducerConnect, tcuEGLStreamProducerConnect, "cuEGLStreamProducerConnect");
LOAD_SYMBOL_OPT(cuEGLStreamProducerDisconnect, tcuEGLStreamProducerDisconnect, "cuEGLStreamProducerDisconnect");
LOAD_SYMBOL_OPT(cuEGLStreamConsumerDisconnect, tcuEGLStreamConsumerDisconnect, "cuEGLStreamConsumerDisconnect");
LOAD_SYMBOL_OPT(cuEGLStreamProducerPresentFrame, tcuEGLStreamProducerPresentFrame, "cuEGLStreamProducerPresentFrame");
LOAD_SYMBOL_OPT(cuEGLStreamProducerReturnFrame, tcuEGLStreamProducerReturnFrame, "cuEGLStreamProducerReturnFrame");
#if defined(_WIN32) || defined(__CYGWIN__)
LOAD_SYMBOL(cuD3D11GetDevice, tcuD3D11GetDevice, "cuD3D11GetDevice");
LOAD_SYMBOL(cuD3D11GetDevices, tcuD3D11GetDevices, "cuD3D11GetDevices");
LOAD_SYMBOL(cuGraphicsD3D11RegisterResource, tcuGraphicsD3D11RegisterResource, "cuGraphicsD3D11RegisterResource");
#endif
GENERIC_LOAD_FUNC_FINALE(cuda);
}
#endif
static inline int cuvid_load_functions(CuvidFunctions **functions, void *logctx)
{
(void)logctx;
GENERIC_LOAD_FUNC_PREAMBLE(CuvidFunctions, cuvid, NVCUVID_LIBNAME);
LOAD_SYMBOL_OPT(cuvidGetDecoderCaps, tcuvidGetDecoderCaps, "cuvidGetDecoderCaps");
LOAD_SYMBOL(cuvidCreateDecoder, tcuvidCreateDecoder, "cuvidCreateDecoder");
LOAD_SYMBOL(cuvidDestroyDecoder, tcuvidDestroyDecoder, "cuvidDestroyDecoder");
LOAD_SYMBOL(cuvidDecodePicture, tcuvidDecodePicture, "cuvidDecodePicture");
LOAD_SYMBOL(cuvidGetDecodeStatus, tcuvidGetDecodeStatus, "cuvidGetDecodeStatus");
LOAD_SYMBOL(cuvidReconfigureDecoder, tcuvidReconfigureDecoder, "cuvidReconfigureDecoder");
#ifdef __CUVID_DEVPTR64
LOAD_SYMBOL(cuvidMapVideoFrame, tcuvidMapVideoFrame, "cuvidMapVideoFrame64");
LOAD_SYMBOL(cuvidUnmapVideoFrame, tcuvidUnmapVideoFrame, "cuvidUnmapVideoFrame64");
#else
LOAD_SYMBOL(cuvidMapVideoFrame, tcuvidMapVideoFrame, "cuvidMapVideoFrame");
LOAD_SYMBOL(cuvidUnmapVideoFrame, tcuvidUnmapVideoFrame, "cuvidUnmapVideoFrame");
#endif
LOAD_SYMBOL(cuvidCtxLockCreate, tcuvidCtxLockCreate, "cuvidCtxLockCreate");
LOAD_SYMBOL(cuvidCtxLockDestroy, tcuvidCtxLockDestroy, "cuvidCtxLockDestroy");
LOAD_SYMBOL(cuvidCtxLock, tcuvidCtxLock, "cuvidCtxLock");
LOAD_SYMBOL(cuvidCtxUnlock, tcuvidCtxUnlock, "cuvidCtxUnlock");
#if !defined(__APPLE__)
LOAD_SYMBOL(cuvidCreateVideoSource, tcuvidCreateVideoSource, "cuvidCreateVideoSource");
LOAD_SYMBOL(cuvidCreateVideoSourceW, tcuvidCreateVideoSourceW, "cuvidCreateVideoSourceW");
LOAD_SYMBOL(cuvidDestroyVideoSource, tcuvidDestroyVideoSource, "cuvidDestroyVideoSource");
LOAD_SYMBOL(cuvidSetVideoSourceState, tcuvidSetVideoSourceState, "cuvidSetVideoSourceState");
LOAD_SYMBOL(cuvidGetVideoSourceState, tcuvidGetVideoSourceState, "cuvidGetVideoSourceState");
LOAD_SYMBOL(cuvidGetSourceVideoFormat, tcuvidGetSourceVideoFormat, "cuvidGetSourceVideoFormat");
LOAD_SYMBOL(cuvidGetSourceAudioFormat, tcuvidGetSourceAudioFormat, "cuvidGetSourceAudioFormat");
#endif
LOAD_SYMBOL(cuvidCreateVideoParser, tcuvidCreateVideoParser, "cuvidCreateVideoParser");
LOAD_SYMBOL(cuvidParseVideoData, tcuvidParseVideoData, "cuvidParseVideoData");
LOAD_SYMBOL(cuvidDestroyVideoParser, tcuvidDestroyVideoParser, "cuvidDestroyVideoParser");
GENERIC_LOAD_FUNC_FINALE(cuvid);
}
static inline int nvenc_load_functions(NvencFunctions **functions, void *logctx)
{
(void)logctx;
GENERIC_LOAD_FUNC_PREAMBLE(NvencFunctions, nvenc, NVENC_LIBNAME);
LOAD_SYMBOL(NvEncodeAPICreateInstance, tNvEncodeAPICreateInstance, "NvEncodeAPICreateInstance");
LOAD_SYMBOL(NvEncodeAPIGetMaxSupportedVersion, tNvEncodeAPIGetMaxSupportedVersion, "NvEncodeAPIGetMaxSupportedVersion");
GENERIC_LOAD_FUNC_FINALE(nvenc);
}
#undef GENERIC_LOAD_FUNC_PREAMBLE
#undef LOAD_LIBRARY
#undef LOAD_SYMBOL
#undef GENERIC_LOAD_FUNC_FINALE
#undef GENERIC_FREE_FUNC
#undef CUDA_LIBNAME
#undef NVCUVID_LIBNAME
#undef NVENC_LIBNAME
#endif

View File

@@ -0,0 +1,499 @@
/*
* This copyright notice applies to this header file only:
*
* Copyright (c) 2010-2023 NVIDIA Corporation
*
* 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.
*/
/********************************************************************************************************************/
//! \file nvcuvid.h
//! NVDECODE API provides video decoding interface to NVIDIA GPU devices.
//! \date 2015-2022
//! This file contains the interface constants, structure definitions and function prototypes.
/********************************************************************************************************************/
#if !defined(__NVCUVID_H__)
#define __NVCUVID_H__
#include "dynlink_cuviddec.h"
#if defined(__cplusplus)
extern "C" {
#endif /* __cplusplus */
#define MAX_CLOCK_TS 3
/***********************************************/
//!
//! High-level helper APIs for video sources
//!
/***********************************************/
typedef void *CUvideosource;
typedef void *CUvideoparser;
typedef long long CUvideotimestamp;
/************************************************************************/
//! \enum cudaVideoState
//! Video source state enums
//! Used in cuvidSetVideoSourceState and cuvidGetVideoSourceState APIs
/************************************************************************/
typedef enum {
cudaVideoState_Error = -1, /**< Error state (invalid source) */
cudaVideoState_Stopped = 0, /**< Source is stopped (or reached end-of-stream) */
cudaVideoState_Started = 1 /**< Source is running and delivering data */
} cudaVideoState;
/************************************************************************/
//! \enum cudaAudioCodec
//! Audio compression enums
//! Used in CUAUDIOFORMAT structure
/************************************************************************/
typedef enum {
cudaAudioCodec_MPEG1=0, /**< MPEG-1 Audio */
cudaAudioCodec_MPEG2, /**< MPEG-2 Audio */
cudaAudioCodec_MP3, /**< MPEG-1 Layer III Audio */
cudaAudioCodec_AC3, /**< Dolby Digital (AC3) Audio */
cudaAudioCodec_LPCM, /**< PCM Audio */
cudaAudioCodec_AAC, /**< AAC Audio */
} cudaAudioCodec;
/************************************************************************/
//! \ingroup STRUCTS
//! \struct HEVCTIMECODESET
//! Used to store Time code extracted from Time code SEI in HEVC codec
/************************************************************************/
typedef struct _HEVCTIMECODESET
{
unsigned int time_offset_value;
unsigned short n_frames;
unsigned char clock_timestamp_flag;
unsigned char units_field_based_flag;
unsigned char counting_type;
unsigned char full_timestamp_flag;
unsigned char discontinuity_flag;
unsigned char cnt_dropped_flag;
unsigned char seconds_value;
unsigned char minutes_value;
unsigned char hours_value;
unsigned char seconds_flag;
unsigned char minutes_flag;
unsigned char hours_flag;
unsigned char time_offset_length;
unsigned char reserved;
} HEVCTIMECODESET;
/************************************************************************/
//! \ingroup STRUCTS
//! \struct HEVCSEITIMECODE
//! Used to extract Time code SEI in HEVC codec
/************************************************************************/
typedef struct _HEVCSEITIMECODE
{
HEVCTIMECODESET time_code_set[MAX_CLOCK_TS];
unsigned char num_clock_ts;
} HEVCSEITIMECODE;
/**********************************************************************************/
//! \ingroup STRUCTS
//! \struct CUSEIMESSAGE;
//! Used in CUVIDSEIMESSAGEINFO structure
/**********************************************************************************/
typedef struct _CUSEIMESSAGE
{
unsigned char sei_message_type; /**< OUT: SEI Message Type */
unsigned char reserved[3];
unsigned int sei_message_size; /**< OUT: SEI Message Size */
} CUSEIMESSAGE;
/************************************************************************************************/
//! \ingroup STRUCTS
//! \struct CUVIDEOFORMAT
//! Video format
//! Used in cuvidGetSourceVideoFormat API
/************************************************************************************************/
typedef struct
{
cudaVideoCodec codec; /**< OUT: Compression format */
/**
* OUT: frame rate = numerator / denominator (for example: 30000/1001)
*/
struct {
/**< OUT: frame rate numerator (0 = unspecified or variable frame rate) */
unsigned int numerator;
/**< OUT: frame rate denominator (0 = unspecified or variable frame rate) */
unsigned int denominator;
} frame_rate;
unsigned char progressive_sequence; /**< OUT: 0=interlaced, 1=progressive */
unsigned char bit_depth_luma_minus8; /**< OUT: high bit depth luma. E.g, 2 for 10-bitdepth, 4 for 12-bitdepth */
unsigned char bit_depth_chroma_minus8; /**< OUT: high bit depth chroma. E.g, 2 for 10-bitdepth, 4 for 12-bitdepth */
unsigned char min_num_decode_surfaces; /**< OUT: Minimum number of decode surfaces to be allocated for correct
decoding. The client can send this value in ulNumDecodeSurfaces
(in CUVIDDECODECREATEINFO structure).
This guarantees correct functionality and optimal video memory
usage but not necessarily the best performance, which depends on
the design of the overall application. The optimal number of
decode surfaces (in terms of performance and memory utilization)
should be decided by experimentation for each application, but it
cannot go below min_num_decode_surfaces.
If this value is used for ulNumDecodeSurfaces then it must be
returned to parser during sequence callback. */
unsigned int coded_width; /**< OUT: coded frame width in pixels */
unsigned int coded_height; /**< OUT: coded frame height in pixels */
/**
* area of the frame that should be displayed
* typical example:
* coded_width = 1920, coded_height = 1088
* display_area = { 0,0,1920,1080 }
*/
struct {
int left; /**< OUT: left position of display rect */
int top; /**< OUT: top position of display rect */
int right; /**< OUT: right position of display rect */
int bottom; /**< OUT: bottom position of display rect */
} display_area;
cudaVideoChromaFormat chroma_format; /**< OUT: Chroma format */
unsigned int bitrate; /**< OUT: video bitrate (bps, 0=unknown) */
/**
* OUT: Display Aspect Ratio = x:y (4:3, 16:9, etc)
*/
struct {
int x;
int y;
} display_aspect_ratio;
/**
* Video Signal Description
* Refer section E.2.1 (VUI parameters semantics) of H264 spec file
*/
struct {
unsigned char video_format : 3; /**< OUT: 0-Component, 1-PAL, 2-NTSC, 3-SECAM, 4-MAC, 5-Unspecified */
unsigned char video_full_range_flag : 1; /**< OUT: indicates the black level and luma and chroma range */
unsigned char reserved_zero_bits : 4; /**< Reserved bits */
unsigned char color_primaries; /**< OUT: chromaticity coordinates of source primaries */
unsigned char transfer_characteristics; /**< OUT: opto-electronic transfer characteristic of the source picture */
unsigned char matrix_coefficients; /**< OUT: used in deriving luma and chroma signals from RGB primaries */
} video_signal_description;
unsigned int seqhdr_data_length; /**< OUT: Additional bytes following (CUVIDEOFORMATEX) */
} CUVIDEOFORMAT;
/****************************************************************/
//! \ingroup STRUCTS
//! \struct CUVIDOPERATINGPOINTINFO
//! Operating point information of scalable bitstream
/****************************************************************/
typedef struct
{
cudaVideoCodec codec;
union
{
struct
{
unsigned char operating_points_cnt;
unsigned char reserved24_bits[3];
unsigned short operating_points_idc[32];
} av1;
unsigned char CodecReserved[1024];
};
} CUVIDOPERATINGPOINTINFO;
/**********************************************************************************/
//! \ingroup STRUCTS
//! \struct CUVIDSEIMESSAGEINFO
//! Used in cuvidParseVideoData API with PFNVIDSEIMSGCALLBACK pfnGetSEIMsg
/**********************************************************************************/
typedef struct _CUVIDSEIMESSAGEINFO
{
void *pSEIData; /**< OUT: SEI Message Data */
CUSEIMESSAGE *pSEIMessage; /**< OUT: SEI Message Info */
unsigned int sei_message_count; /**< OUT: SEI Message Count */
unsigned int picIdx; /**< OUT: SEI Message Pic Index */
} CUVIDSEIMESSAGEINFO;
/****************************************************************/
//! \ingroup STRUCTS
//! \struct CUVIDAV1SEQHDR
//! AV1 specific sequence header information
/****************************************************************/
typedef struct {
unsigned int max_width;
unsigned int max_height;
unsigned char reserved[1016];
} CUVIDAV1SEQHDR;
/****************************************************************/
//! \ingroup STRUCTS
//! \struct CUVIDEOFORMATEX
//! Video format including raw sequence header information
//! Used in cuvidGetSourceVideoFormat API
/****************************************************************/
typedef struct
{
CUVIDEOFORMAT format; /**< OUT: CUVIDEOFORMAT structure */
union {
CUVIDAV1SEQHDR av1;
unsigned char raw_seqhdr_data[1024]; /**< OUT: Sequence header data */
};
} CUVIDEOFORMATEX;
/****************************************************************/
//! \ingroup STRUCTS
//! \struct CUAUDIOFORMAT
//! Audio formats
//! Used in cuvidGetSourceAudioFormat API
/****************************************************************/
typedef struct
{
cudaAudioCodec codec; /**< OUT: Compression format */
unsigned int channels; /**< OUT: number of audio channels */
unsigned int samplespersec; /**< OUT: sampling frequency */
unsigned int bitrate; /**< OUT: For uncompressed, can also be used to determine bits per sample */
unsigned int reserved1; /**< Reserved for future use */
unsigned int reserved2; /**< Reserved for future use */
} CUAUDIOFORMAT;
/***************************************************************/
//! \enum CUvideopacketflags
//! Data packet flags
//! Used in CUVIDSOURCEDATAPACKET structure
/***************************************************************/
typedef enum {
CUVID_PKT_ENDOFSTREAM = 0x01, /**< Set when this is the last packet for this stream */
CUVID_PKT_TIMESTAMP = 0x02, /**< Timestamp is valid */
CUVID_PKT_DISCONTINUITY = 0x04, /**< Set when a discontinuity has to be signalled */
CUVID_PKT_ENDOFPICTURE = 0x08, /**< Set when the packet contains exactly one frame or one field */
CUVID_PKT_NOTIFY_EOS = 0x10, /**< If this flag is set along with CUVID_PKT_ENDOFSTREAM, an additional (dummy)
display callback will be invoked with null value of CUVIDPARSERDISPINFO which
should be interpreted as end of the stream. */
} CUvideopacketflags;
/*****************************************************************************/
//! \ingroup STRUCTS
//! \struct CUVIDSOURCEDATAPACKET
//! Data Packet
//! Used in cuvidParseVideoData API
//! IN for cuvidParseVideoData
/*****************************************************************************/
typedef struct _CUVIDSOURCEDATAPACKET
{
tcu_ulong flags; /**< IN: Combination of CUVID_PKT_XXX flags */
tcu_ulong payload_size; /**< IN: number of bytes in the payload (may be zero if EOS flag is set) */
const unsigned char *payload; /**< IN: Pointer to packet payload data (may be NULL if EOS flag is set) */
CUvideotimestamp timestamp; /**< IN: Presentation time stamp (10MHz clock), only valid if
CUVID_PKT_TIMESTAMP flag is set */
} CUVIDSOURCEDATAPACKET;
// Callback for packet delivery
typedef int (CUDAAPI *PFNVIDSOURCECALLBACK)(void *, CUVIDSOURCEDATAPACKET *);
/**************************************************************************************************************************/
//! \ingroup STRUCTS
//! \struct CUVIDSOURCEPARAMS
//! Describes parameters needed in cuvidCreateVideoSource API
//! NVDECODE API is intended for HW accelerated video decoding so CUvideosource doesn't have audio demuxer for all supported
//! containers. It's recommended to clients to use their own or third party demuxer if audio support is needed.
/**************************************************************************************************************************/
typedef struct _CUVIDSOURCEPARAMS
{
unsigned int ulClockRate; /**< IN: Time stamp units in Hz (0=default=10000000Hz) */
unsigned int bAnnexb : 1; /**< IN: AV1 annexB stream */
unsigned int uReserved : 31; /**< Reserved for future use - set to zero */
unsigned int uReserved1[6]; /**< Reserved for future use - set to zero */
void *pUserData; /**< IN: User private data passed in to the data handlers */
PFNVIDSOURCECALLBACK pfnVideoDataHandler; /**< IN: Called to deliver video packets */
PFNVIDSOURCECALLBACK pfnAudioDataHandler; /**< IN: Called to deliver audio packets. */
void *pvReserved2[8]; /**< Reserved for future use - set to NULL */
} CUVIDSOURCEPARAMS;
/**********************************************/
//! \ingroup ENUMS
//! \enum CUvideosourceformat_flags
//! CUvideosourceformat_flags
//! Used in cuvidGetSourceVideoFormat API
/**********************************************/
typedef enum {
CUVID_FMT_EXTFORMATINFO = 0x100 /**< Return extended format structure (CUVIDEOFORMATEX) */
} CUvideosourceformat_flags;
#if !defined(__APPLE__)
/***************************************************************************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidCreateVideoSource(CUvideosource *pObj, const char *pszFileName, CUVIDSOURCEPARAMS *pParams)
//! Create CUvideosource object. CUvideosource spawns demultiplexer thread that provides two callbacks:
//! pfnVideoDataHandler() and pfnAudioDataHandler()
//! NVDECODE API is intended for HW accelerated video decoding so CUvideosource doesn't have audio demuxer for all supported
//! containers. It's recommended to clients to use their own or third party demuxer if audio support is needed.
/***************************************************************************************************************************/
typedef CUresult CUDAAPI tcuvidCreateVideoSource(CUvideosource *pObj, const char *pszFileName, CUVIDSOURCEPARAMS *pParams);
/***************************************************************************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidCreateVideoSourceW(CUvideosource *pObj, const wchar_t *pwszFileName, CUVIDSOURCEPARAMS *pParams)
//! Create video source
/***************************************************************************************************************************/
typedef CUresult CUDAAPI tcuvidCreateVideoSourceW(CUvideosource *pObj, const wchar_t *pwszFileName, CUVIDSOURCEPARAMS *pParams);
/********************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidDestroyVideoSource(CUvideosource obj)
//! Destroy video source
/********************************************************************/
typedef CUresult CUDAAPI tcuvidDestroyVideoSource(CUvideosource obj);
/******************************************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidSetVideoSourceState(CUvideosource obj, cudaVideoState state)
//! Set video source state to:
//! cudaVideoState_Started - to signal the source to run and deliver data
//! cudaVideoState_Stopped - to stop the source from delivering the data
//! cudaVideoState_Error - invalid source
/******************************************************************************************/
typedef CUresult CUDAAPI tcuvidSetVideoSourceState(CUvideosource obj, cudaVideoState state);
/******************************************************************************************/
//! \ingroup FUNCTS
//! \fn cudaVideoState CUDAAPI cuvidGetVideoSourceState(CUvideosource obj)
//! Get video source state
//! Returns:
//! cudaVideoState_Started - if Source is running and delivering data
//! cudaVideoState_Stopped - if Source is stopped or reached end-of-stream
//! cudaVideoState_Error - if Source is in error state
/******************************************************************************************/
typedef cudaVideoState CUDAAPI tcuvidGetVideoSourceState(CUvideosource obj);
/******************************************************************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidGetSourceVideoFormat(CUvideosource obj, CUVIDEOFORMAT *pvidfmt, unsigned int flags)
//! Gets video source format in pvidfmt, flags is set to combination of CUvideosourceformat_flags as per requirement
/******************************************************************************************************************/
typedef CUresult CUDAAPI tcuvidGetSourceVideoFormat(CUvideosource obj, CUVIDEOFORMAT *pvidfmt, unsigned int flags);
/**************************************************************************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidGetSourceAudioFormat(CUvideosource obj, CUAUDIOFORMAT *paudfmt, unsigned int flags)
//! Get audio source format
//! NVDECODE API is intended for HW accelerated video decoding so CUvideosource doesn't have audio demuxer for all supported
//! containers. It's recommended to clients to use their own or third party demuxer if audio support is needed.
/**************************************************************************************************************************/
typedef CUresult CUDAAPI tcuvidGetSourceAudioFormat(CUvideosource obj, CUAUDIOFORMAT *paudfmt, unsigned int flags);
#endif
/**********************************************************************************/
//! \ingroup STRUCTS
//! \struct CUVIDPARSERDISPINFO
//! Used in cuvidParseVideoData API with PFNVIDDISPLAYCALLBACK pfnDisplayPicture
/**********************************************************************************/
typedef struct _CUVIDPARSERDISPINFO
{
int picture_index; /**< OUT: Index of the current picture */
int progressive_frame; /**< OUT: 1 if progressive frame; 0 otherwise */
int top_field_first; /**< OUT: 1 if top field is displayed first; 0 otherwise */
int repeat_first_field; /**< OUT: Number of additional fields (1=ivtc, 2=frame doubling, 4=frame tripling,
-1=unpaired field) */
CUvideotimestamp timestamp; /**< OUT: Presentation time stamp */
} CUVIDPARSERDISPINFO;
/***********************************************************************************************************************/
//! Parser callbacks
//! The parser will call these synchronously from within cuvidParseVideoData(), whenever there is sequence change or a picture
//! is ready to be decoded and/or displayed. First argument in functions is "void *pUserData" member of structure CUVIDSOURCEPARAMS
//! Return values from these callbacks are interpreted as below. If the callbacks return failure, it will be propagated by
//! cuvidParseVideoData() to the application.
//! Parser picks default operating point as 0 and outputAllLayers flag as 0 if PFNVIDOPPOINTCALLBACK is not set or return value is
//! -1 or invalid operating point.
//! PFNVIDSEQUENCECALLBACK : 0: fail, 1: succeeded, > 1: override dpb size of parser (set by CUVIDPARSERPARAMS::ulMaxNumDecodeSurfaces
//! while creating parser)
//! PFNVIDDECODECALLBACK : 0: fail, >=1: succeeded
//! PFNVIDDISPLAYCALLBACK : 0: fail, >=1: succeeded
//! PFNVIDOPPOINTCALLBACK : <0: fail, >=0: succeeded (bit 0-9: OperatingPoint, bit 10-10: outputAllLayers, bit 11-30: reserved)
//! PFNVIDSEIMSGCALLBACK : 0: fail, >=1: succeeded
/***********************************************************************************************************************/
typedef int (CUDAAPI *PFNVIDSEQUENCECALLBACK)(void *, CUVIDEOFORMAT *);
typedef int (CUDAAPI *PFNVIDDECODECALLBACK)(void *, CUVIDPICPARAMS *);
typedef int (CUDAAPI *PFNVIDDISPLAYCALLBACK)(void *, CUVIDPARSERDISPINFO *);
typedef int (CUDAAPI *PFNVIDOPPOINTCALLBACK)(void *, CUVIDOPERATINGPOINTINFO*);
typedef int (CUDAAPI *PFNVIDSEIMSGCALLBACK) (void *, CUVIDSEIMESSAGEINFO *);
/**************************************/
//! \ingroup STRUCTS
//! \struct CUVIDPARSERPARAMS
//! Used in cuvidCreateVideoParser API
/**************************************/
typedef struct _CUVIDPARSERPARAMS
{
cudaVideoCodec CodecType; /**< IN: cudaVideoCodec_XXX */
unsigned int ulMaxNumDecodeSurfaces; /**< IN: Max # of decode surfaces (parser will cycle through these) */
unsigned int ulClockRate; /**< IN: Timestamp units in Hz (0=default=10000000Hz) */
unsigned int ulErrorThreshold; /**< IN: % Error threshold (0-100) for calling pfnDecodePicture (100=always
IN: call pfnDecodePicture even if picture bitstream is fully corrupted) */
unsigned int ulMaxDisplayDelay; /**< IN: Max display queue delay (improves pipelining of decode with display)
0=no delay (recommended values: 2..4) */
unsigned int bAnnexb : 1; /**< IN: AV1 annexB stream */
unsigned int uReserved : 31; /**< Reserved for future use - set to zero */
unsigned int uReserved1[4]; /**< IN: Reserved for future use - set to 0 */
void *pUserData; /**< IN: User data for callbacks */
PFNVIDSEQUENCECALLBACK pfnSequenceCallback; /**< IN: Called before decoding frames and/or whenever there is a fmt change */
PFNVIDDECODECALLBACK pfnDecodePicture; /**< IN: Called when a picture is ready to be decoded (decode order) */
PFNVIDDISPLAYCALLBACK pfnDisplayPicture; /**< IN: Called whenever a picture is ready to be displayed (display order) */
PFNVIDOPPOINTCALLBACK pfnGetOperatingPoint; /**< IN: Called from AV1 sequence header to get operating point of a AV1
scalable bitstream */
PFNVIDSEIMSGCALLBACK pfnGetSEIMsg; /**< IN: Called when all SEI messages are parsed for particular frame */
void *pvReserved2[5]; /**< Reserved for future use - set to NULL */
CUVIDEOFORMATEX *pExtVideoInfo; /**< IN: [Optional] sequence header data from system layer */
} CUVIDPARSERPARAMS;
/************************************************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidCreateVideoParser(CUvideoparser *pObj, CUVIDPARSERPARAMS *pParams)
//! Create video parser object and initialize
/************************************************************************************************/
typedef CUresult CUDAAPI tcuvidCreateVideoParser(CUvideoparser *pObj, CUVIDPARSERPARAMS *pParams);
/************************************************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidParseVideoData(CUvideoparser obj, CUVIDSOURCEDATAPACKET *pPacket)
//! Parse the video data from source data packet in pPacket
//! Extracts parameter sets like SPS, PPS, bitstream etc. from pPacket and
//! calls back pfnDecodePicture with CUVIDPICPARAMS data for kicking of HW decoding
//! calls back pfnSequenceCallback with CUVIDEOFORMAT data for initial sequence header or when
//! the decoder encounters a video format change
//! calls back pfnDisplayPicture with CUVIDPARSERDISPINFO data to display a video frame
/************************************************************************************************/
typedef CUresult CUDAAPI tcuvidParseVideoData(CUvideoparser obj, CUVIDSOURCEDATAPACKET *pPacket);
/************************************************************************************************/
//! \ingroup FUNCTS
//! \fn CUresult CUDAAPI cuvidDestroyVideoParser(CUvideoparser obj)
//! Destroy the video parser
/************************************************************************************************/
typedef CUresult CUDAAPI tcuvidDestroyVideoParser(CUvideoparser obj);
/**********************************************************************************************/
#if defined(__cplusplus)
}
#endif /* __cplusplus */
#endif // __NVCUVID_H__

File diff suppressed because it is too large Load Diff