mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-12-25 09:40:37 +00:00
Code Issues 1 Projects Releases Wiki Activity
gstreamer/sys/nvcodec/gstcudaloader.h

168 lines
6.8 KiB
C
Raw Normal View History

nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
/* GStreamer
* Copyright (C) 2019 Seungha Yang <seungha.yang@navercorp.com>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library 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
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public
* License along with this library; if not, write to the
* Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
* Boston, MA 02110-1301, USA.
*/
#ifndef __GST_CUDA_LOADER_H__
#define __GST_CUDA_LOADER_H__
nvcodec: cuda.h only needs glib.h, not gst.h Just a nitpick. Also, force the compiler to use our stub header instead of searching for it in the include paths.
2020-01-20 09:40:51 +00:00
#include "stub/cuda.h"
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
#include <gst/gst.h>
G_BEGIN_DECLS
gboolean gst_cuda_load_library (void);
/* cuda.h */
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuInit (unsigned int Flags);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGetErrorName (CUresult error,
const char **pStr);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGetErrorString (CUresult error,
const char **pStr);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuCtxCreate (CUcontext * pctx,
unsigned int flags,
CUdevice dev);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuCtxDestroy (CUcontext ctx);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuCtxPopCurrent (CUcontext * pctx);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuCtxPushCurrent (CUcontext ctx);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
cudacontext: Enable direct CUDA memory access over multiple GPUs If each device context can access each other, enable peer access for better interoperability. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1633>
2019-08-30 08:19:44 +00:00
CUresult CUDAAPI CuCtxEnablePeerAccess (CUcontext peerContext,
unsigned int Flags);
CUresult CUDAAPI CuCtxDisablePeerAccess (CUcontext peerContext);
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGraphicsMapResources (unsigned int count,
CUgraphicsResource * resources,
CUstream hStream);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGraphicsUnmapResources (unsigned int count,
CUgraphicsResource * resources,
CUstream hStream);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGraphicsSubResourceGetMappedArray (CUarray * pArray,
CUgraphicsResource resource,
unsigned int arrayIndex,
unsigned int mipLevel);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGraphicsResourceGetMappedPointer (CUdeviceptr * pDevPtr,
size_t * pSize,
CUgraphicsResource resource);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGraphicsUnregisterResource (CUgraphicsResource resource);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuMemAlloc (CUdeviceptr * dptr,
unsigned int bytesize);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuMemAllocPitch (CUdeviceptr * dptr,
size_t * pPitch,
size_t WidthInBytes,
size_t Height,
unsigned int ElementSizeBytes);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Add CUDA specific memory and bufferpool Introducing CUDA buffer pool with generic CUDA memory support. Likewise GL memory, any elements which are able to access CUDA device memory directly can map this CUDA memory without upload/download overhead via the "GST_MAP_CUDA" map flag. Also usual GstMemory map/unmap is also possible with internal staging memory. For staging, CUDA Host allocated memory is used (see CuMemAllocHost API). The memory is allowing system access but has lower overhead during GPU upload/download than normal system memory. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1633>
2019-08-19 09:02:56 +00:00
CUresult CUDAAPI CuMemAllocHost (void **pp,
unsigned int bytesize);
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuMemcpy2D (const CUDA_MEMCPY2D * pCopy);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuMemcpy2DAsync (const CUDA_MEMCPY2D *pCopy, CUstream hStream);
nvdec: Make OpenGL dependency optional By adding system memory support for nvdec, both en/decoder in the nvcodec plugin are able to be usable regardless of OpenGL dependency. Besides, the direct use of system memory might have less overhead than OpenGL memory depending on use cases. (e.g., transcoding using S/W encoder)
2019-07-19 13:46:01 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuMemFree (CUdeviceptr dptr);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Add CUDA specific memory and bufferpool Introducing CUDA buffer pool with generic CUDA memory support. Likewise GL memory, any elements which are able to access CUDA device memory directly can map this CUDA memory without upload/download overhead via the "GST_MAP_CUDA" map flag. Also usual GstMemory map/unmap is also possible with internal staging memory. For staging, CUDA Host allocated memory is used (see CuMemAllocHost API). The memory is allowing system access but has lower overhead during GPU upload/download than normal system memory. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1633>
2019-08-19 09:02:56 +00:00
CUresult CUDAAPI CuMemFreeHost (void *p);
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuStreamCreate (CUstream *phStream,
unsigned int Flags);
nvdec: Don't use default CUDA stream Async CUDA operation with default stream (NULL CUstream) is not much beneficial than blocking operation since all CUDA operations which belong to the CUDA context will be synchronized with the default stream's operation. Note that CUDA stream will share all resources of the corresponding CUDA context but which can help parallel operation similar to the relation between thread and process
2019-08-18 13:07:38 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuStreamDestroy (CUstream hStream);
nvdec: Don't use default CUDA stream Async CUDA operation with default stream (NULL CUstream) is not much beneficial than blocking operation since all CUDA operations which belong to the CUDA context will be synchronized with the default stream's operation. Note that CUDA stream will share all resources of the corresponding CUDA context but which can help parallel operation similar to the relation between thread and process
2019-08-18 13:07:38 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuStreamSynchronize (CUstream hStream);
nvdec: Make OpenGL dependency optional By adding system memory support for nvdec, both en/decoder in the nvcodec plugin are able to be usable regardless of OpenGL dependency. Besides, the direct use of system memory might have less overhead than OpenGL memory depending on use cases. (e.g., transcoding using S/W encoder)
2019-07-19 13:46:01 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuDeviceGet (CUdevice * device,
int ordinal);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuDeviceGetCount (int *count);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuDeviceGetName (char *name,
int len,
CUdevice dev);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuDeviceGetAttribute (int *pi,
CUdevice_attribute attrib,
CUdevice dev);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
cudacontext: Enable direct CUDA memory access over multiple GPUs If each device context can access each other, enable peer access for better interoperability. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1633>
2019-08-30 08:19:44 +00:00
CUresult CUDAAPI CuDeviceCanAccessPeer (int *canAccessPeer,
CUdevice dev,
CUdevice peerDev);
nvcodec: Add generic CUDA video convert object Introducing generic video convert object similar to video-converter but using CUDA. Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1633>
2019-10-16 13:42:39 +00:00
CUresult CUDAAPI CuDriverGetVersion (int * driverVersion);
CUresult CUDAAPI CuModuleLoadData (CUmodule* module,
const void *image);
CUresult CUDAAPI CuModuleUnload (CUmodule module);
CUresult CUDAAPI CuModuleGetFunction (CUfunction* hfunc,
CUmodule hmod,
const char* name);
CUresult CUDAAPI CuTexObjectCreate (CUtexObject *pTexObject,
const CUDA_RESOURCE_DESC *pResDesc,
const CUDA_TEXTURE_DESC *pTexDesc,
const CUDA_RESOURCE_VIEW_DESC *pResViewDesc);
CUresult CUDAAPI CuTexObjectDestroy (CUtexObject texObject);
CUresult CUDAAPI CuLaunchKernel (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);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
/* cudaGL.h */
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGraphicsGLRegisterImage (CUgraphicsResource * pCudaResource,
unsigned int image,
unsigned int target,
unsigned int Flags);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGraphicsGLRegisterBuffer (CUgraphicsResource * pCudaResource,
unsigned int buffer,
unsigned int Flags);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
nvcodec: Fix crash on 32-bit Windows We weren't using the correct calling convention when calling CUDA and CUVID APIs. `CUDAAPI` is `__stdcall` on Windows. This was working fine on x64 because `__stdcall` is ignored and there's no special calling convention. However, on x86, we need to use `__stdcall`.
2020-01-20 09:52:26 +00:00
CUresult CUDAAPI CuGraphicsResourceSetMapFlags (CUgraphicsResource resource,
unsigned int flags);
cudautils: Add GstCudaGraphicsResource structure for better openGL interoperability Introduce GstCudaGraphicsResource structure to represent registered CUDA graphics resources and to enable sharing the information among nvdec and nvenc. This structure can reduce the number of resource registration which cause high overhead.
2019-08-17 08:45:44 +00:00
nvdec: Add fallback for CUDA/OpenGL interop failure It happens when local OpenGL context belongs to non-nvidia GPU.
2020-03-15 10:20:47 +00:00
CUresult CUDAAPI CuGLGetDevices (unsigned int * pCudaDeviceCount,
CUdevice * pCudaDevices,
unsigned int cudaDeviceCount,
CUGLDeviceList deviceList);
nvdec,nvenc: Port to dynamic library loading ... and put them into new nvcodec plugin. * nvcodec plugin Now each nvenc and nvdec element is moved to be a part of nvcodec plugin for better interoperability. Additionally, cuda runtime API header dependencies (i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed. Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have been used except for some defined types. However, those types could be replaced with other types which were defined by "cuda.h". * dynamic library loading CUDA library will be opened with g_module_open() instead of build-time linking. On Windows, nvcuda.dll is installed to system path by CUDA Toolkit installer, and on *nix, user should ensure that libcuda.so.1 can be loadable (i.e., via LD_LIBRARY_PATH or default dlopen path) Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows is removed.
2019-05-17 13:27:50 +00:00
G_END_DECLS
#endif /* __GST_CUDA_LOADER_H__ */
Reference in a new issue Copy permalink