Commit graph

2801 commits

Author SHA1 Message Date
Seungha Yang
a1ba38cd64 d3d12: Add compositor element
Adding d3d12compositor element. d3d12compositor will build GPU commands
asynchronously and each command is serialized at final render stage.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5875>
2024-01-02 13:02:47 +00:00
Seungha Yang
996346f8d3 d3d12converter: Don't map output buffer with write flag
Conversion will happen when constructed command list is executed,
not by converter element. Thus this object should not map output buffer
with write flag which will result in error if multiple threads
are building commands for the same output target frame.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5875>
2024-01-02 13:02:47 +00:00
Seungha Yang
bffbf0d951 d3d12converter: Add support for texture upload
If buffer is not a d3d12 memory or allocated by other device,
upload to internal d3d12 memory

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5875>
2024-01-02 13:02:47 +00:00
Seungha Yang
f4fe17d8d2 d3d12converter: Add support for blending
Create new PSO if blend state update is required

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5875>
2024-01-02 13:02:47 +00:00
Seungha Yang
7c701058ed d3d12: Add testsrc element
Adding testsrc element with d2d interop support via d3d11on12

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5875>
2024-01-02 13:02:46 +00:00
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
7d210af46c docs: Update Windows plugins cache
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
2f091e7118 d3d12: Add video sink element
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
660f2d7d27 d3d12: Add convert element
Implement converter object with convert element

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
4198bd6932 d3d12: Add helper object for fence operation
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
eeb57061d2 d3d12: Enable plugin only for Windows8 or newer
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
3e49d6a75f d3d12: Define more formats
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
48cfca413d d3d12: Add header containing core features
... and include the single header instead of listing many ones

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
41d8a12649 d3d12: Remove unused methods
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
5ce2a7f64f d3d12bufferpool: Don't pre-allocate memory for size calculation
Unlike d3d11, we can know CPU accessible memory layout without
allocation

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
d316356bf2 d3d12memory: Add alloc_wrapped() method
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
3308a976bd d3d12: Add upload element
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
e4f794cbdd d3d12bufferpool: Wait fence before reusing buffer
Buffer can be released without waiting fence for previous commands

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:39:00 +00:00
Seungha Yang
1c5bba4b6b d3d12decoder: Remove ID3D12Device4 interface requirement
Old OS may not support the interface. And allow 11_0 feature level
hardware.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:38:59 +00:00
Seungha Yang
4e5d4a45a3 d3d12decoder: Reduce the number of resource barriers
Single barrier per texture is sufficient in case of array-of-textures.
Avoid unnecessary decay barriers

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:38:59 +00:00
Seungha Yang
e6bdb0458c d3d12decoder: Use flexible task queue
Instead of using fixed size command allocator array, make it
resizable.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:38:59 +00:00
Seungha Yang
efc023e76e d3d12: Rework command scheduling
* Use single fence object per queue and remove GstD3D12Fence
  implementation
* Add a helper method for texture copy
* Run background thread and release unused resource from the thread

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:38:59 +00:00
Seungha Yang
6d7d9291c3 d3d12: Add resource pool objects
Adding pool objects for command list, command allocator, and descriptor
heap

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:38:59 +00:00
Seungha Yang
93458c0155 d3d12memory: Add more SRV/RTV getter methods
Adding a method so that memory object can create SRV/RTV
on external descriptor heap. And remove unused methods

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:38:59 +00:00
Seungha Yang
2578f234dd d3d12: Remove d3d11 dependency
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5870>
2023-12-29 14:38:59 +00:00
Olivier Crête
e19428a802 onnx: Update build instructions to use onnx-runtime 0.16.3
This synchronizes it with the meson.build file

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5861>
2023-12-22 14:43:23 -05:00
Víctor Manuel Jáquez Leal
fdfd51397b vajpegdec: only support progressive mjpeg streams
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Víctor Manuel Jáquez Leal
ec4054e637 jpegparse: always forward pixel-aspect-ratio
And by default it's 1/1

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Víctor Manuel Jáquez Leal
ae55ebfdf2 jpegparse: fix endianness
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Víctor Manuel Jáquez Leal
470b90a14a jpegparse: on app0 use fourcc rather than strings
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Víctor Manuel Jáquez Leal
367deef132 jpegparse: dump whole failed app marker
Instead of just log the parsed string. It's better for debugging.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Víctor Manuel Jáquez Leal
658b5b4399 jpegparse: don't trigger message for failed com marker
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Víctor Manuel Jáquez Leal
497890f678 jpegparse: refactor renegotation
Instead of comparing caps, just signal when renegotation is needed.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Víctor Manuel Jáquez Leal
03d51244c0 jpegparse: pass colorimetry in caps
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Víctor Manuel Jáquez Leal
20b636ee9a jpegparse: add support interlaced mjpeg
Interlaced MJPEG is a big hack. Most of the streams we've found are from old
AVID tools. There are two methods to detect interlaced stream: the container
offers a height bigger (or double) than the image's height in SOF. The other
is from a APP0 marker.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5838>
2023-12-22 04:38:06 +00:00
Aaron Boxer
e2ee207367 onnx: add README outlining install and test instructions
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5816>
2023-12-20 14:48:41 -05:00
Marek Vasut
5f3d4215a0 v4l2codecs: Switch gst_codec_picture_ts_ns() to gst_util_uint64_scale_int()
Instead of plain multiplication, use gst_util_uint64_scale_int()
to achieve the same effect with additional checks.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Marek Vasut
3335af0efe v4l2codecs: Deduplicate picture frame number to timestamp in ns
Add macro which converts picture frame number to suitable timestamp in
nanoseconds for use in V4L2 VB2 buffer lookup. Since multiple codecs do
the same operation and almost all got it wrong, do it in one place so it
can be fixed in one place again, if needed.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Marek Vasut
552a171671 vp9decoder: Simplify gst_v4l2_codecs_vp9_dec_fill_refs()
In case reference_frames is NULL, return outright. Remove the
duplicate check from subsequent conditionals. No functional change.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Marek Vasut
4560cdff5c h265decoder: Align wraparound fix
Instead of casting GST_CODEC_PICTURE_FRAME_NUMBER (ref_pic) to u64,
use 1000ULL which is also u64 . This only aligns the behavior here
with '*decoder: Fix multiplication wraparound' commits.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Marek Vasut
7725aa7d77 h264decoder: Align wraparound fix
Instead of casting GST_CODEC_PICTURE_FRAME_NUMBER (ref_pic) to u64,
use 1000ULL which is also u64 . This only aligns the behavior here
with '*decoder: Fix multiplication wraparound' commits.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Marek Vasut
3cbf09d0c9 mpeg2decoder: Fix multiplication wraparound
The GstMpeg2Picture system_frame_number is guint32, constant 1000 is guint32,
GstV4l2CodecMpeg2Dec *_ref_ts multiplication result is u64 .

```
u64 result = (u32)((u32)system_frame_number * (u32)1000);
```
behaves the same as
```
u64 result = (u32)(((u32)system_frame_number * (u32)1000) & 0xffffffff);
```

so in case `system_frame_number > 4294967295 / 1000`, the `result` will
wrap around. Since the `result` is really used as a cookie used to look
up V4L2 buffers related to the currently decoded frame, this wraparound
leads to visible corruption during MPEG2 decoding. At 30 FPS this occurs
after cca. 40 hours of playback .

Fix this by changing the 1000 from u32 to u64, i.e.:
```
u64 result = (u64)((u32)system_frame_number * (u64)1000ULL);
```
this way, the wraparound is prevented and the correct cookie is used.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Marek Vasut
50fb6f8c02 av1decoder: Fix multiplication wraparound
The GstAV1Picture system_frame_number is guint32, constant 1000 is guint32,
GstV4l2CodecAV1Dec v4l2_av1_frame.*_frame_ts multiplication result is u64 .

```
u64 result = (u32)((u32)system_frame_number * (u32)1000);
```
behaves the same as
```
u64 result = (u32)(((u32)system_frame_number * (u32)1000) & 0xffffffff);
```

so in case `system_frame_number > 4294967295 / 1000`, the `result` will
wrap around. Since the `result` is really used as a cookie used to look
up V4L2 buffers related to the currently decoded frame, this wraparound
leads to visible corruption during AV1 decoding. At 30 FPS this occurs
after cca. 40 hours of playback .

Fix this by changing the 1000 from u32 to u64, i.e.:
```
u64 result = (u64)((u32)system_frame_number * (u64)1000ULL);
```
this way, the wraparound is prevented and the correct cookie is used.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Marek Vasut
6f74818a07 vp9decoder: Fix multiplication wraparound
The GstVp9Picture system_frame_number is guint32, constant 1000 is guint32,
GstV4l2CodecVp9Dec v4l2_vp9_frame.*_frame_ts multiplication result is u64 .

```
u64 result = (u32)((u32)system_frame_number * (u32)1000);
```
behaves the same as
```
u64 result = (u32)(((u32)system_frame_number * (u32)1000) & 0xffffffff);
```

so in case `system_frame_number > 4294967295 / 1000`, the `result` will
wrap around. Since the `result` is really used as a cookie used to look
up V4L2 buffers related to the currently decoded frame, this wraparound
leads to visible corruption during VP9 decoding. At 30 FPS this occurs
after cca. 40 hours of playback .

Fix this by changing the 1000 from u32 to u64, i.e.:
```
u64 result = (u64)((u32)system_frame_number * (u64)1000ULL);
```
this way, the wraparound is prevented and the correct cookie is used.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Marek Vasut
15c9a6caa9 vp8decoder: Fix multiplication wraparound
The GstVp8Picture system_frame_number is guint32, constant 1000 is guint32,
GstV4l2CodecVp8Dec v4l2_vp8_frame.*_frame_ts multiplication result is u64 .

```
u64 result = (u32)((u32)system_frame_number * (u32)1000);
```
behaves the same as
```
u64 result = (u32)(((u32)system_frame_number * (u32)1000) & 0xffffffff);
```

so in case `system_frame_number > 4294967295 / 1000`, the `result` will
wrap around. Since the `result` is really used as a cookie used to look
up V4L2 buffers related to the currently decoded frame, this wraparound
leads to visible corruption during VP8 decoding. At 30 FPS this occurs
after cca. 40 hours of playback .

Fix this by changing the 1000 from u32 to u64, i.e.:
```
u64 result = (u64)((u32)system_frame_number * (u64)1000ULL);
```
this way, the wraparound is prevented and the correct cookie is used.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5791>
2023-12-20 18:47:39 +00:00
Mengkejiergeli Ba
141cd38715 h264parse: Remove dead code
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5833>
2023-12-20 13:03:25 +00:00
Sebastian Dröge
2e86fb691a video-format: Fix format order once again
RGBA should be before RBGA. Both the Python script and the gstreamer-rs
tests agree on that, but somehow this is not caught by the CI.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5837>
2023-12-20 05:33:43 +00:00
Olivier Crête
f5d5f2cf1a meta: Add API to register metas in two steps
And also remove the specific registration APIs for
serializable meta.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5830>
2023-12-19 22:41:31 +00:00
Olivier Crête
2a9c4e3270 meta: Move the clear operation to its own vfunc
Some transforms always assumed that the transformation was some kind
of copy. So adding a "clear" operation didn't work out in practice.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5830>
2023-12-19 22:41:31 +00:00
Philippe Normand
61d3f5c8e0 play: Include pipeline dump in error details structure
This can be useful for debugging purposes. It can't be done on application side
because the on_error callback tears down the pipeline.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/5828>
2023-12-18 20:13:38 +00:00