Commit graph

53 commits

Author SHA1 Message Date
Seungha Yang
abe1f5044d cuda: Prefer CUBIN over PTX
System installed NVRTC library might be newer version than
driver, then generate PTX can be incompatible with the driver.
Instead of the intermediate code PTX, use actual assembly code
directly.

Fixes: https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/3108
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5639>
2024-01-02 10:10:09 +00:00
Seungha Yang
91e0c3aafa cuda: Use d3d11 token data for interop data
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5409>
2023-09-29 12:36:01 +00:00
Seungha Yang
c818906236 cuda: Add support for I420_12LE format
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5409>
2023-09-29 12:36:01 +00:00
Seungha Yang
a80f542f66 cuda: Add support for P012_LE and Y444/GBR high bitdepth formats
Adding P012, Y444_10, Y444_12, GBR_10, GBR_12 and GBR_16 formats support

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5375>
2023-09-23 13:12:55 +00:00
Seungha Yang
dc2fd997a2 cuda: Add workaround for gir build
ERROR:../girepository/girparser.c:343:state_switch:
  assertion failed: (ctx->state != newstate)
Bail out! ERROR:../girepository/girparser.c:343:state_switch:
  assertion failed: (ctx->state != newstate)

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4510>
2023-08-14 13:41:01 +00:00
Seungha Yang
7b1e4d6051 cudabufferpool: Add support for virtual memory
Configure malloc or mmap allocator depending on config option

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4510>
2023-08-14 13:41:01 +00:00
Seungha Yang
547b13c68f cudacontext: Add memory allocation related properties
Adding "virtual-memory" and "os-handle" properties. New properties
will be used to query device's capability

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4510>
2023-08-14 13:41:01 +00:00
Seungha Yang
2f506e8ddc cudamemory: Add support for virtual memory in pool allocator
Adding new memory pool allocator for virtual memory

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4510>
2023-08-14 13:41:01 +00:00
Seungha Yang
194fd8bb82 cudamemory: Add support for virtual memory management
Adding new CUDA memory allocation methods

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4510>
2023-08-14 13:41:01 +00:00
Seungha Yang
a712a768a4 cuda: Load virtual memory management and IPC API symbols
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4510>
2023-08-14 13:41:01 +00:00
Seungha Yang
de749fa356 cuda: Introduce GST_CUDA_CRITICAL_ERRORS env to abort on critical error
Adding GST_CUDA_CRITICAL_ERRORS env variable so that program can be
terminated on unrecoverable error.

Example)
GST_CUDA_CRITICAL_ERRORS=2,700 gst-launch-1.0 ...

In this example, CUDA_ERROR_OUT_OF_MEMORY(2) and
CUDA_ERROR_ILLEGAL_ADDRESS(700) are registered as critical error
and program will be aborted on those errors

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4729>
2023-06-18 16:44:43 +00:00
Seungha Yang
58b166453d cuda: Move cuda debug helper function to .cpp
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4729>
2023-06-18 16:44:43 +00:00
Seungha Yang
1aa9e74aaf cudadownload: Always download CUDA memory if it's bound to decoder
Decoder bounded CUDA memory is allocated by driver and the pool size
is fixed. Since we don't know how many buffers would be held by
downstream non-CUDA element, we should download such CUDA memory
and release it back to decoder.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4810>
2023-06-08 22:27:06 +00:00
Seungha Yang
4ed3c46de7 cudamemory: Fix for semi planar YUV memory size decision
UV plan of the semi planar format requires only half of Y plane size

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4502>
2023-04-27 20:55:53 +00:00
Thibault Saunier
b14e675a27 gir: Checkout all .gir files and check that they are updated on the CI
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3010>
2023-04-22 09:32:32 -04:00
Alicia Boya García
85b6625150 cudaloader: Initialize logging category
gstcudaloader.cpp defines GST_DEBUG_CATEGORY (gst_cudaloader_debug);
but it wasn't initializing it anywhere.

This caused the following error to be logged by gst-plugin-scanner when
libcuda.so.1/nvcuda.dll couldn't be loaded, e.g. in systems without
CUDA:

(gst-plugin-scanner:39618): GStreamer-CRITICAL **: 14:40:22.346:
gst_debug_log_full_valist: assertion 'category != NULL' failed

This patch fixes the bug by initializing the category in
gst_cuda_load_library_once_func() before any logging occurs.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/4154>
2023-03-14 03:07:50 +00:00
Seungha Yang
319f5f0760 cuda: Link libatomic if needed
Looks like C++ does not pull it automatically

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3997>
2023-02-21 16:00:32 +00:00
Seungha Yang
59f359eb99 cuda: Rename macro HAVE_NVCODEC_GST_GL -> HAVE_CUDA_GST_GL
... and always use #ifdef instead of #if

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3992>
2023-02-20 01:49:31 +09:00
Seungha Yang
ff3120a38c cudamemory, d3d11memory: Add memory_{get,set}_token_data() methods
Similar to GstMiniObject qdata but new methods will use int64
token value and per object lock, instead of GQuark with global
mutex in qdata

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3884>
2023-02-16 17:49:54 +00:00
Seungha Yang
f6defc0c5b cudamemory: Add gst_cuda_allocator_alloc_wrapped() method
... so that application can pass already allocated CUDA memory

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3884>
2023-02-16 17:49:54 +00:00
Seungha Yang
e77e6fd4a7 cudamemory: Skip sync if no I/O operation happend on free()
Synchronization for unused memory is not required

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3884>
2023-02-16 17:49:54 +00:00
Seungha Yang
f44cac1c9f cudamemory: Make CUtexObject object reusable
Create and hold CUtexObject objects in GstCudaMemory so that it can
be reusable

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3884>
2023-02-16 17:49:54 +00:00
Seungha Yang
992406cf4f cuda, nvcodec: Make GstD3D11 dependency mandatory
GstD3D11 build-time dependencies should be always available on Windows already
and runtime dependencies as well, since required external
(non-GStreamer) depends are all system DLLs

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3884>
2023-02-16 17:49:54 +00:00
Seungha Yang
f212bd901b cuda: Port to C++
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3884>
2023-02-16 17:49:53 +00:00
Seungha Yang
090d50e1a0 tests: Add CUDA memory allocator test
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3629>
2023-02-03 15:27:43 +00:00
Seungha Yang
7a8bb85523 cudaupload, cudadownload: Update for shared CUDA stream
Use CUDA stream of memory if exists

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3629>
2023-02-03 15:27:43 +00:00
Seungha Yang
d409c35367 cudabufferpool: Add support for CUDA stream use in memory
* Use GstCudaPoolAllocator
* Pass configured GstCudaStream object to allocator

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3629>
2023-02-03 15:27:42 +00:00
Seungha Yang
30d06e03c2 cudamemory: Make GstCudaStream-aware
This will be used for CUDA stream sharing.

* Adding GstCudaPoolAllocator object. The pool allocator will
  control synchronization of allocated memory objects.
* Modify gst_cuda_allocator_alloc() API so that caller can specify/set
  GstCudaStream object for the newly allocated memory.
* GST_CUDA_MEMORY_TRANSFER_NEED_SYNC flag is added in addition to
  existing GST_CUDA_MEMORY_TRANSFER_NEED_{UPLOAD,DOWNLOAD}.
  The flag indicates that any GPU command queued in the CUDA stream
  may not be finished yet, and caller should take care of the
  synchronization.
  The flag is controlled by GstCudaMemory object if the memory holds
  GstCudaStream. (Otherwise, GstCudaMemory will do synchronization
  as before this commit). Specifically, GstCudaMemory object will set
  the new flag automatically when memory is mapped with
  (GST_MAP_CUDA | GST_MAP_WRITE) flags. Caller will need to unset
  the flag via GST_MEMORY_FLAG_UNSET() if it's already synchronized
  by client code.
* gst_cuda_memory_sync() helper function is added to perform synchronization
* Why not use CUevent object to keep track of synchronization status?
  CUDA provides fence-like interface already via CUevent object,
  but cuEventRecord/cuEventQuery APIs are not zero-cost operations.
  Instead, in this version, the status is tracked by using map and
  object flags.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3629>
2023-02-03 15:27:42 +00:00
Seungha Yang
9eaae61a44 cudamemory: Allow nullptr allocator object
The GstCudaAllocator object doesn't hold any device object.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3629>
2023-02-03 15:27:42 +00:00
Seungha Yang
a7c54ebc06 cuda: Add GstCudaStream object
Wrap CUstream handle with GstCudaStream to make it ref-counted
object. This GstCudaStream object will be used later for
CUDA stream sharing

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3629>
2023-02-03 15:27:42 +00:00
Seungha Yang
661b5f60c6 cuda: Provide single header include entry point
Add "gstcuda.h" header file

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3629>
2023-02-03 15:27:42 +00:00
Seungha Yang
a874c9f2d4 cuda: Fix runtime compiler library loading on Windows
The cuda is a part of GPU driver but runtime compiler is a part of
cuda toolkit, which means the version number can be different.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3545>
2022-12-09 19:24:08 +09:00
Seungha Yang
e6585c89ea cuda: Hide memory copy util function
The method was intended to be used by only cudaupload/download elements
and not ready to be a part of public API

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3545>
2022-12-08 19:52:23 +00:00
Seungha Yang
c1efa9ac4b cudaupload,cudadownload: Add support for planar 8bits RGB formats
Defines RGBP, BGRP, and GBR formats, which have the same memory
layout as already supported Y444. And defines GBRA format as well
which needs just one additional alpha plane

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3389>
2022-11-15 16:25:44 +00:00
Seungha Yang
9c4c8c3932 cudacontext: Store texture alignment
it was missed in the previous refactoring

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3389>
2022-11-15 16:25:44 +00:00
Sebastian Dröge
73557fe093 cuda: Add/fix various annotations
And fix a memory leak when creating a CUDA context fails.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3194>
2022-10-18 13:51:17 +03:00
Seungha Yang
219bb769a2 cuda: Remove GST_CUDA_HAS_D3D define from header
... and fix d3d11 specific enum type name

GST_CUDA_HAS_D3D is a build time define which indicates whether
GstD3D11 library is available or not, but DirectX SDK headers
must be available on the build system already.

Expose Direct3D related symbols if the build target is Windows
(i.e., if G_OS_WIN32 is defined)

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3018>
2022-09-13 16:40:51 +00:00
Thibault Saunier
bc9c1e3956 meson: Namespace the plugins_doc_dep/libraries variables
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2970>
2022-09-01 21:17:35 +00:00
Thibault Saunier
c971a6d71f cuda: meson: Fix generating GstCuda gir file against gstreamer-full
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2970>
2022-09-01 21:17:35 +00:00
Seungha Yang
75e8f80999 cuda: Hide runtime compiler related header and symbols
That's already abstracted via gst_cuda_nvrtc_compile() method
and therefore, we do not need to expose such symbols yet.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2884>
2022-08-18 05:20:08 +09:00
Seungha Yang
bcacfc7ced cuda: Protect gst_cuda_ensure_element_context() with lock
Serialize the order of context setup since it can be called
from multiple threads

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2666>
2022-06-30 12:38:08 +00:00
Seungha Yang
ce82c0570e cuda: Add YUY2 and UYVY formats for GL interop
Those formats are not supported conversion formats by CUDA convert/scale elements
but would be useful for GL <-> CUDA interop use case.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2691>
2022-06-30 11:17:15 +00:00
Corentin Damman
8678fc999a cudamemory: add support for planar 4:2:2 YUV formats
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2269>
2022-06-30 07:30:09 +00:00
Seungha Yang
9d0f792b66 meson: cuda: Add missing override_dependency()
Required for GstCuda library to be consumed via "gstreamer-cuda-1.0"
dependency.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2539>
2022-06-02 19:14:12 +00:00
Seungha Yang
dc1d5f798f cuda: Fix uninitialized debug category
Make sure debug category init

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2542>
2022-06-02 18:33:30 +00:00
Thibault Saunier
dd585ecfcc cuda: Fix introspection warnings
Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1183

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2344>
2022-05-02 18:30:22 +00:00
Seungha Yang
1fb3804408 Revert "cuda: Fix GstCudaContext refcount bug"
This reverts commit df7ee0e215.

We should not unref it since gst_message_new_have_context()
takes ownership of the GstContext.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2342>
2022-05-02 17:19:13 +00:00
Myles Inglis
df7ee0e215 cuda: Fix GstCudaContext refcount bug
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2334>
2022-04-29 22:22:25 +00:00
Seungha Yang
9cd6f9d51e cuda: Correct CUDA device id field in GstContext structure
device id is unsigned int, not signed.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2311>
2022-04-29 21:21:04 +00:00
Seungha Yang
205a23f456 nvcodec: Stop building for non-{Windows,Linux} target
The other platforms, specifically macOS have not been supported already
because this plugin loads so or dll. Moreover, NVIDIA dropped
support for macOS as of CUDA 11.0. See also
https://developer.nvidia.com/nvidia-cuda-toolkit-11_0_0-developer-tools-mac-hosts

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2283>
2022-04-25 15:50:52 +00:00