gstreamer/sys/nvcodec/gstcudaloader.c
Seungha Yang d16a9237c3 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>
2020-10-16 15:56:49 +00:00

460 lines
14 KiB
C

/* 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.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "gstcudaloader.h"
#include <gmodule.h>
GST_DEBUG_CATEGORY_EXTERN (gst_nvcodec_debug);
#define GST_CAT_DEFAULT gst_nvcodec_debug
#ifndef G_OS_WIN32
#define CUDA_LIBNAME "libcuda.so.1"
#else
#define CUDA_LIBNAME "nvcuda.dll"
#endif
#define LOAD_SYMBOL(name,func) G_STMT_START { \
if (!g_module_symbol (module, G_STRINGIFY (name), (gpointer *) &vtable->func)) { \
GST_ERROR ("Failed to load '%s' from %s, %s", G_STRINGIFY (name), filename, g_module_error()); \
goto error; \
} \
} G_STMT_END;
typedef struct _GstNvCodecCudaVTable
{
gboolean loaded;
CUresult (CUDAAPI * CuInit) (unsigned int Flags);
CUresult (CUDAAPI * CuGetErrorName) (CUresult error, const char **pStr);
CUresult (CUDAAPI * CuGetErrorString) (CUresult error, const char **pStr);
CUresult (CUDAAPI * CuCtxCreate) (CUcontext * pctx, unsigned int flags,
CUdevice dev);
CUresult (CUDAAPI * CuCtxDestroy) (CUcontext ctx);
CUresult (CUDAAPI * CuCtxPopCurrent) (CUcontext * pctx);
CUresult (CUDAAPI * CuCtxPushCurrent) (CUcontext ctx);
CUresult (CUDAAPI * CuCtxEnablePeerAccess) (CUcontext peerContext,
unsigned int Flags);
CUresult (CUDAAPI * CuCtxDisablePeerAccess) (CUcontext peerContext);
CUresult (CUDAAPI * CuGraphicsMapResources) (unsigned int count,
CUgraphicsResource * resources, CUstream hStream);
CUresult (CUDAAPI * CuGraphicsUnmapResources) (unsigned int count,
CUgraphicsResource * resources, CUstream hStream);
CUresult (CUDAAPI * CuGraphicsSubResourceGetMappedArray) (CUarray * pArray,
CUgraphicsResource resource, unsigned int arrayIndex,
unsigned int mipLevel);
CUresult (CUDAAPI * CuGraphicsResourceGetMappedPointer) (CUdeviceptr *
pDevPtr, size_t * pSize, CUgraphicsResource resource);
CUresult (CUDAAPI *
CuGraphicsUnregisterResource) (CUgraphicsResource resource);
CUresult (CUDAAPI * CuMemAlloc) (CUdeviceptr * dptr, unsigned int bytesize);
CUresult (CUDAAPI * CuMemAllocPitch) (CUdeviceptr * dptr, size_t * pPitch,
size_t WidthInBytes, size_t Height, unsigned int ElementSizeBytes);
CUresult (CUDAAPI * CuMemAllocHost) (void **pp, unsigned int bytesize);
CUresult (CUDAAPI * CuMemcpy2D) (const CUDA_MEMCPY2D * pCopy);
CUresult (CUDAAPI * CuMemcpy2DAsync) (const CUDA_MEMCPY2D * pCopy,
CUstream hStream);
CUresult (CUDAAPI * CuMemFree) (CUdeviceptr dptr);
CUresult (CUDAAPI * CuMemFreeHost) (void *p);
CUresult (CUDAAPI * CuStreamCreate) (CUstream * phStream,
unsigned int Flags);
CUresult (CUDAAPI * CuStreamDestroy) (CUstream hStream);
CUresult (CUDAAPI * CuStreamSynchronize) (CUstream hStream);
CUresult (CUDAAPI * CuDeviceGet) (CUdevice * device, int ordinal);
CUresult (CUDAAPI * CuDeviceGetCount) (int *count);
CUresult (CUDAAPI * CuDeviceGetName) (char *name, int len, CUdevice dev);
CUresult (CUDAAPI * CuDeviceGetAttribute) (int *pi,
CUdevice_attribute attrib, CUdevice dev);
CUresult (CUDAAPI * CuDeviceCanAccessPeer) (int *canAccessPeer,
CUdevice dev, CUdevice peerDev);
CUresult (CUDAAPI * CuGraphicsGLRegisterImage) (CUgraphicsResource *
pCudaResource, unsigned int image, unsigned int target,
unsigned int Flags);
CUresult (CUDAAPI * CuGraphicsGLRegisterBuffer) (CUgraphicsResource *
pCudaResource, unsigned int buffer, unsigned int Flags);
CUresult (CUDAAPI *
CuGraphicsResourceSetMapFlags) (CUgraphicsResource resource,
unsigned int flags);
CUresult (CUDAAPI * CuGLGetDevices) (unsigned int *pCudaDeviceCount,
CUdevice * pCudaDevices, unsigned int cudaDeviceCount,
CUGLDeviceList deviceList);
} GstNvCodecCudaVTable;
static GstNvCodecCudaVTable gst_cuda_vtable = { 0, };
gboolean
gst_cuda_load_library (void)
{
GModule *module;
const gchar *filename = CUDA_LIBNAME;
GstNvCodecCudaVTable *vtable;
if (gst_cuda_vtable.loaded)
return TRUE;
module = g_module_open (filename, G_MODULE_BIND_LAZY);
if (module == NULL) {
GST_WARNING ("Could not open library %s, %s", filename, g_module_error ());
return FALSE;
}
vtable = &gst_cuda_vtable;
/* cuda.h */
LOAD_SYMBOL (cuInit, CuInit);
LOAD_SYMBOL (cuGetErrorName, CuGetErrorName);
LOAD_SYMBOL (cuGetErrorString, CuGetErrorString);
LOAD_SYMBOL (cuCtxCreate, CuCtxCreate);
LOAD_SYMBOL (cuCtxDestroy, CuCtxDestroy);
LOAD_SYMBOL (cuCtxPopCurrent, CuCtxPopCurrent);
LOAD_SYMBOL (cuCtxPushCurrent, CuCtxPushCurrent);
LOAD_SYMBOL (cuCtxEnablePeerAccess, CuCtxEnablePeerAccess);
LOAD_SYMBOL (cuCtxDisablePeerAccess, CuCtxDisablePeerAccess);
LOAD_SYMBOL (cuGraphicsMapResources, CuGraphicsMapResources);
LOAD_SYMBOL (cuGraphicsUnmapResources, CuGraphicsUnmapResources);
LOAD_SYMBOL (cuGraphicsSubResourceGetMappedArray,
CuGraphicsSubResourceGetMappedArray);
LOAD_SYMBOL (cuGraphicsResourceGetMappedPointer,
CuGraphicsResourceGetMappedPointer);
LOAD_SYMBOL (cuGraphicsUnregisterResource, CuGraphicsUnregisterResource);
LOAD_SYMBOL (cuMemAlloc, CuMemAlloc);
LOAD_SYMBOL (cuMemAllocPitch, CuMemAllocPitch);
LOAD_SYMBOL (cuMemAllocHost, CuMemAllocHost);
LOAD_SYMBOL (cuMemcpy2D, CuMemcpy2D);
LOAD_SYMBOL (cuMemcpy2DAsync, CuMemcpy2DAsync);
LOAD_SYMBOL (cuMemFree, CuMemFree);
LOAD_SYMBOL (cuMemFreeHost, CuMemFreeHost);
LOAD_SYMBOL (cuStreamCreate, CuStreamCreate);
LOAD_SYMBOL (cuStreamDestroy, CuStreamDestroy);
LOAD_SYMBOL (cuStreamSynchronize, CuStreamSynchronize);
LOAD_SYMBOL (cuDeviceGet, CuDeviceGet);
LOAD_SYMBOL (cuDeviceGetCount, CuDeviceGetCount);
LOAD_SYMBOL (cuDeviceGetName, CuDeviceGetName);
LOAD_SYMBOL (cuDeviceGetAttribute, CuDeviceGetAttribute);
LOAD_SYMBOL (cuDeviceCanAccessPeer, CuDeviceCanAccessPeer);
/* cudaGL.h */
LOAD_SYMBOL (cuGraphicsGLRegisterImage, CuGraphicsGLRegisterImage);
LOAD_SYMBOL (cuGraphicsGLRegisterBuffer, CuGraphicsGLRegisterBuffer);
LOAD_SYMBOL (cuGraphicsResourceSetMapFlags, CuGraphicsResourceSetMapFlags);
LOAD_SYMBOL (cuGLGetDevices, CuGLGetDevices);
vtable->loaded = TRUE;
return TRUE;
error:
g_module_close (module);
return FALSE;
}
CUresult CUDAAPI
CuInit (unsigned int Flags)
{
g_assert (gst_cuda_vtable.CuInit != NULL);
return gst_cuda_vtable.CuInit (Flags);
}
CUresult CUDAAPI
CuGetErrorName (CUresult error, const char **pStr)
{
g_assert (gst_cuda_vtable.CuGetErrorName != NULL);
return gst_cuda_vtable.CuGetErrorName (error, pStr);
}
CUresult CUDAAPI
CuGetErrorString (CUresult error, const char **pStr)
{
g_assert (gst_cuda_vtable.CuGetErrorString != NULL);
return gst_cuda_vtable.CuGetErrorString (error, pStr);
}
CUresult CUDAAPI
CuCtxCreate (CUcontext * pctx, unsigned int flags, CUdevice dev)
{
g_assert (gst_cuda_vtable.CuCtxCreate != NULL);
return gst_cuda_vtable.CuCtxCreate (pctx, flags, dev);
}
CUresult CUDAAPI
CuCtxDestroy (CUcontext ctx)
{
g_assert (gst_cuda_vtable.CuCtxDestroy != NULL);
return gst_cuda_vtable.CuCtxDestroy (ctx);
}
CUresult CUDAAPI
CuCtxPopCurrent (CUcontext * pctx)
{
g_assert (gst_cuda_vtable.CuCtxPopCurrent != NULL);
return gst_cuda_vtable.CuCtxPopCurrent (pctx);
}
CUresult CUDAAPI
CuCtxPushCurrent (CUcontext ctx)
{
g_assert (gst_cuda_vtable.CuCtxPushCurrent != NULL);
return gst_cuda_vtable.CuCtxPushCurrent (ctx);
}
CUresult CUDAAPI
CuCtxEnablePeerAccess (CUcontext peerContext, unsigned int Flags)
{
g_assert (gst_cuda_vtable.CuCtxEnablePeerAccess != NULL);
return gst_cuda_vtable.CuCtxEnablePeerAccess (peerContext, Flags);
}
CUresult CUDAAPI
CuCtxDisablePeerAccess (CUcontext peerContext)
{
g_assert (gst_cuda_vtable.CuCtxDisablePeerAccess != NULL);
return gst_cuda_vtable.CuCtxDisablePeerAccess (peerContext);
}
CUresult CUDAAPI
CuGraphicsMapResources (unsigned int count, CUgraphicsResource * resources,
CUstream hStream)
{
g_assert (gst_cuda_vtable.CuGraphicsMapResources != NULL);
return gst_cuda_vtable.CuGraphicsMapResources (count, resources, hStream);
}
CUresult CUDAAPI
CuGraphicsUnmapResources (unsigned int count, CUgraphicsResource * resources,
CUstream hStream)
{
g_assert (gst_cuda_vtable.CuGraphicsUnmapResources != NULL);
return gst_cuda_vtable.CuGraphicsUnmapResources (count, resources, hStream);
}
CUresult CUDAAPI
CuGraphicsSubResourceGetMappedArray (CUarray * pArray,
CUgraphicsResource resource, unsigned int arrayIndex, unsigned int mipLevel)
{
g_assert (gst_cuda_vtable.CuGraphicsSubResourceGetMappedArray != NULL);
return gst_cuda_vtable.CuGraphicsSubResourceGetMappedArray (pArray, resource,
arrayIndex, mipLevel);
}
CUresult CUDAAPI
CuGraphicsResourceGetMappedPointer (CUdeviceptr * pDevPtr, size_t * pSize,
CUgraphicsResource resource)
{
g_assert (gst_cuda_vtable.CuGraphicsResourceGetMappedPointer != NULL);
return gst_cuda_vtable.CuGraphicsResourceGetMappedPointer (pDevPtr, pSize,
resource);
}
CUresult CUDAAPI
CuGraphicsUnregisterResource (CUgraphicsResource resource)
{
g_assert (gst_cuda_vtable.CuGraphicsUnregisterResource != NULL);
return gst_cuda_vtable.CuGraphicsUnregisterResource (resource);
}
CUresult CUDAAPI
CuMemAlloc (CUdeviceptr * dptr, unsigned int bytesize)
{
g_assert (gst_cuda_vtable.CuMemAlloc != NULL);
return gst_cuda_vtable.CuMemAlloc (dptr, bytesize);
}
CUresult CUDAAPI
CuMemAllocPitch (CUdeviceptr * dptr, size_t * pPitch, size_t WidthInBytes,
size_t Height, unsigned int ElementSizeBytes)
{
g_assert (gst_cuda_vtable.CuMemAllocPitch != NULL);
return gst_cuda_vtable.CuMemAllocPitch (dptr, pPitch, WidthInBytes, Height,
ElementSizeBytes);
}
CUresult CUDAAPI
CuMemAllocHost (void **pp, unsigned int bytesize)
{
g_assert (gst_cuda_vtable.CuMemAllocHost != NULL);
return gst_cuda_vtable.CuMemAllocHost (pp, bytesize);
}
CUresult CUDAAPI
CuMemcpy2D (const CUDA_MEMCPY2D * pCopy)
{
g_assert (gst_cuda_vtable.CuMemcpy2D != NULL);
return gst_cuda_vtable.CuMemcpy2D (pCopy);
}
CUresult CUDAAPI
CuMemcpy2DAsync (const CUDA_MEMCPY2D * pCopy, CUstream hStream)
{
g_assert (gst_cuda_vtable.CuMemcpy2DAsync != NULL);
return gst_cuda_vtable.CuMemcpy2DAsync (pCopy, hStream);
}
CUresult CUDAAPI
CuMemFree (CUdeviceptr dptr)
{
g_assert (gst_cuda_vtable.CuMemFree != NULL);
return gst_cuda_vtable.CuMemFree (dptr);
}
CUresult CUDAAPI
CuMemFreeHost (void *p)
{
g_assert (gst_cuda_vtable.CuMemFreeHost != NULL);
return gst_cuda_vtable.CuMemFreeHost (p);
}
CUresult CUDAAPI
CuStreamCreate (CUstream * phStream, unsigned int Flags)
{
g_assert (gst_cuda_vtable.CuStreamCreate != NULL);
return gst_cuda_vtable.CuStreamCreate (phStream, Flags);
}
CUresult CUDAAPI
CuStreamDestroy (CUstream hStream)
{
g_assert (gst_cuda_vtable.CuStreamDestroy != NULL);
return gst_cuda_vtable.CuStreamDestroy (hStream);
}
CUresult CUDAAPI
CuStreamSynchronize (CUstream hStream)
{
g_assert (gst_cuda_vtable.CuStreamSynchronize != NULL);
return gst_cuda_vtable.CuStreamSynchronize (hStream);
}
CUresult CUDAAPI
CuDeviceGet (CUdevice * device, int ordinal)
{
g_assert (gst_cuda_vtable.CuDeviceGet != NULL);
return gst_cuda_vtable.CuDeviceGet (device, ordinal);
}
CUresult CUDAAPI
CuDeviceGetCount (int *count)
{
g_assert (gst_cuda_vtable.CuDeviceGetCount != NULL);
return gst_cuda_vtable.CuDeviceGetCount (count);
}
CUresult CUDAAPI
CuDeviceGetName (char *name, int len, CUdevice dev)
{
g_assert (gst_cuda_vtable.CuDeviceGetName != NULL);
return gst_cuda_vtable.CuDeviceGetName (name, len, dev);
}
CUresult CUDAAPI
CuDeviceGetAttribute (int *pi, CUdevice_attribute attrib, CUdevice dev)
{
g_assert (gst_cuda_vtable.CuDeviceGetAttribute != NULL);
return gst_cuda_vtable.CuDeviceGetAttribute (pi, attrib, dev);
}
CUresult CUDAAPI
CuDeviceCanAccessPeer (int *canAccessPeer, CUdevice dev, CUdevice peerDev)
{
g_assert (gst_cuda_vtable.CuDeviceCanAccessPeer != NULL);
return gst_cuda_vtable.CuDeviceCanAccessPeer (canAccessPeer, dev, peerDev);
}
/* cudaGL.h */
CUresult CUDAAPI
CuGraphicsGLRegisterImage (CUgraphicsResource * pCudaResource,
unsigned int image, unsigned int target, unsigned int Flags)
{
g_assert (gst_cuda_vtable.CuGraphicsGLRegisterImage != NULL);
return gst_cuda_vtable.CuGraphicsGLRegisterImage (pCudaResource, image,
target, Flags);
}
CUresult CUDAAPI
CuGraphicsGLRegisterBuffer (CUgraphicsResource * pCudaResource,
unsigned int buffer, unsigned int Flags)
{
g_assert (gst_cuda_vtable.CuGraphicsGLRegisterBuffer != NULL);
return gst_cuda_vtable.CuGraphicsGLRegisterBuffer (pCudaResource, buffer,
Flags);
}
CUresult CUDAAPI
CuGraphicsResourceSetMapFlags (CUgraphicsResource resource, unsigned int flags)
{
g_assert (gst_cuda_vtable.CuGraphicsResourceSetMapFlags != NULL);
return gst_cuda_vtable.CuGraphicsResourceSetMapFlags (resource, flags);
}
CUresult CUDAAPI
CuGLGetDevices (unsigned int *pCudaDeviceCount, CUdevice * pCudaDevices,
unsigned int cudaDeviceCount, CUGLDeviceList deviceList)
{
g_assert (gst_cuda_vtable.CuGLGetDevices != NULL);
return gst_cuda_vtable.CuGLGetDevices (pCudaDeviceCount, pCudaDevices,
cudaDeviceCount, deviceList);
}