summaryrefslogtreecommitdiffstats
path: root/src/video_core/command_classes (follow)
Commit message (Collapse)AuthorAgeFilesLines
* NVDRV: Further refactors and eliminate old code.Fernando Sahmkow2022-10-061-29/+0
|
* VideoCore: Refactor syncing.Fernando Sahmkow2022-10-0617-3011/+0
|
* common: Change semantics of UNREACHABLE to unconditionally crashLiam2022-06-142-2/+2
|
* codecs/vp9: Use u8 for norm and map lutsMorph2022-05-131-4/+4
| | | | Using this smaller type saves 1536 bytes in the compiled executable.
* chore: add missing SPDX tagsAndrea Pappacoda2022-04-284-76/+8
| | | | Follow-up to 99ceb03a1cfcf35968cab589ea188a8c406cda52
* general: Convert source file copyright comments over to SPDXMorph2022-04-2314-42/+28
| | | | | This formats all copyright comments according to SPDX formatting guidelines. Additionally, this resolves the remaining GPLv2 only licensed files by relicensing them to GPLv2.0-or-later.
* Merge pull request #8066 from ameerj/gpu-decode-fixesFernando S2022-04-011-14/+21
|\ | | | | codec: Fix GPU decoder detection on Windows and a memory leak
| * codec: Plug GPU decoder memory leakameerj2022-03-221-0/+2
| |
| * codec: Disable HW_FRAMES method check on Windowsameerj2022-03-221-14/+19
| | | | | | | | | | | | It was reported that this method causes crashes on certain Linux decoding backends, hence the check to avoid it. This subsequently caused Windows GPU decoders to never be selected and always fall back to CPU decoding, disable the check on Windows for now.
* | video_core: Reduce unused includesameerj2022-03-195-6/+1
|/
* video_core: constify AVCodec for ffmpeg >= 5.0Jan Beich2022-01-201-1/+1
| | | | | | src/video_core/command_classes/codecs/codec.cpp:177:16: error: assigning to 'AVCodec *' from 'const AVCodec *' discards qualifiers av_codec = avcodec_find_decoder(codec); ^~~~~~~~~~~~~~~~~~~~~~~~~~~
* video_core/codecs: re-enable VAAPI/VDPAU on BSDs after 72aa418b0b41Jan Beich2021-12-181-1/+1
|
* video_core/codecs: (re-spin) refactor ffmpeg searching and handlingliushuyu2021-12-161-0/+6
|
* Revert "video_core/codecs: refactor ffmpeg searching and handling in cmake"bunnei2021-12-151-6/+0
|
* CI: fix CI on Linuxliushuyu2021-12-141-3/+0
|
* video_core/codecs: skip decoders that use hw frames ...liushuyu2021-12-141-0/+9
| | | | | ... this would resolve some edge-cases where multiple devices are present and ffmpeg is unable to auto-supply the hw surfaces
* video_core/codecs: more fixes for VAAPI detection ...liushuyu2021-12-031-63/+25
| | | | | * skip impersonated VAAPI implementaions ("imposter detection") * place VAAPI priority below CUDA/NVDEC/CUVID
* video_core/codec: address commentsliushuyu2021-12-031-8/+12
|
* video_core/codecs: more robust ffmpeg hwdecoder selection logicliushuyu2021-12-031-10/+27
|
* video_core/codec: address commentsliushuyu2021-11-251-17/+11
|
* video_core/codecs: fix multiple decoding issues on Linux ...liushuyu2021-11-251-2/+47
| | | | | | | | | | * when someone installed Intel video drivers on an AMD system, the decoder will select the Intel VA-API decoding driver and yuzu will crash due to incorrect driver selection; the fix will check if the currently about-to-use driver is loaded in the kernel * when using NVIDIA driver on Linux with a ffmpeg that does not have CUDA capability enabled, the decoder will crash; the fix simply making the decoder prefers the VDPAU driver over CUDA on Linux
* codes: Rename ComposeFrameHeader to ComposeFrameameerj2021-11-137-14/+14
| | | | These functions were composing the entire frame, not just the headers. Rename to more accurately describe them.
* vp8: Implement header compositionameerj2021-11-134-6/+90
| | | | Enables frame decoding with FFmpeg
* codecs: Add VP8 codec classameerj2021-11-138-20/+88
|
* Merge pull request #7157 from ameerj/vic-surface-sizeMorph2021-10-141-16/+15
|\ | | | | vic: Use the minimum of surface/frame dimensions when writing the final frame to the GPU
| * vic: Use the minimum of surface/frame dimensions when writing the final frame to the GPUameerj2021-10-111-16/+15
| | | | | | | | Addresses possible buffer overflow behavior.
* | Merge pull request #7109 from vonchenplus/fix_h264_max__reference_num_errorAmeer J2021-10-121-1/+6
|\ \ | |/ |/| h264: Use max allowed max_num_ref_frames when using CPU decoding
| * h264: Use max allowed max_num_ref_frames when using CPU decodingFeng Chen2021-10-101-1/+6
| |
* | vic: Allow surface to be higher than frameValeri2021-10-091-2/+3
| | | | | | | | | | Touhou Genso Wanderer Lotus Labyrinth R decodes 1920x1080 videos into 1920x1088 surface. Only allow mismatch for height, since larger width would result in increasingly offset rows and somewhat defeat entire purpose of this check.
* | vic: Avoid memory corruption when multiple streams with different dimensions are decodedameerj2021-10-081-0/+9
| | | | | | | | This is a work around to avoid buffer overflow errors until multi channel/multi stream decoding is supported.
* | vic: Refactor frame writing methodsameerj2021-10-072-138/+146
| |
* | vic: Implement RGBX frame formatameerj2021-10-072-3/+15
|/
* codec: Add missing <string_view> includeMorph2021-09-111-0/+1
|
* Merge pull request #6846 from ameerj/nvdec-gpu-decodeFernando S2021-09-113-95/+141
|\ | | | | nvdec: Add GPU video decoding for all capable drivers and platforms
| * h264: Lower max_num_ref_framesameerj2021-08-161-1/+2
| | | | | | | | GPU decoding seems to be more picky when it comes to the maximum number of reference frames.
| * configure_graphics: Add GPU nvdec decoding as an optionameerj2021-08-161-1/+5
| | | | | | | | | | | | Some system configurations may see visual regressions or lower performance using GPU decoding compared to CPU decoding. This setting provides the option for users to specify their decoding preference. Co-Authored-By: yzct12345 <87620833+yzct12345@users.noreply.github.com>
| * codec: Improve libav memory alloc and cleanupameerj2021-08-162-14/+19
| |
| * codec: Fallback to CPU decoding if no compatible GPU format is foundameerj2021-08-162-22/+32
| |
| * codec: Replace deprecated av_init_packet usageameerj2021-08-121-9/+13
| |
| * nvdec: Implement GPU accelerated decoding for all platformsameerj2021-08-122-70/+92
| | | | | | | | Supplements the VAAPI intel gpu decoder by implementing the D3D11VA decoder for Windows, and CUVID/VDPAU for Nvidia and AMD on drivers linux respectively.
* | vp9_types: Minor refactor of VP9 info structs.ameerj2021-08-261-32/+29
| |
* | vp9_types: Remove unused Vp9PictureInfo membersameerj2021-08-262-24/+1
| |
* | Merge pull request #6838 from ameerj/sws-alignbunnei2021-08-121-3/+2
|\ \ | |/ |/| vic: Specify sws_scale height stride.
| * vic: Specify sws_scale height stride.ameerj2021-08-101-3/+2
| | | | | | | | Silences a sws_scale runtime warning about unaligned strides.
* | vp9: Ensure the first frame is completeameerj2021-08-082-3/+3
|/ | | | Silences a runtime error due to the first frame missing the frame data, and being set to hidden despite being a key-frame.
* nvdec: Better logging for unimplemented codecsameerj2021-08-071-1/+1
|
* Merge pull request #6799 from ameerj/vp9-fixesbunnei2021-08-073-106/+48
|\ | | | | nvdec: Fix VP9 reference frame refreshes
| * vp9: Cleanup unused variablesameerj2021-08-073-58/+17
| | | | | | | | | | With reference frames refreshes fix, we no longer need to buffer two frames in advance. We can also remove other unused or otherwise unneeded variables.
| * vp9: Fix reference frame refreshesameerj2021-08-072-48/+31
| | | | | | | | This resolves the artifacting when decoding VP9 streams.
* | nvdec: Implement VA-API hardware video acceleration (#6713)yzct123452021-08-044-72/+170
|/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * nvdec: VA-API * Verify formatting * Forgot a semicolon for Windows * Clarify comment about AV_PIX_FMT_NV12 * Fix assert log spam from missing negation * vic: Remove forgotten debug code * Address lioncash's review * Mention VA-API is Intel/AMD * Address v1993's review * Hopefully fix CMakeLists style this time * vic: Improve cache locality * vic: Fix off-by-one error * codec: Async * codec: Forgot the GetValue() * nvdec: Address ameerj's review * codec: Fallback to CPU without VA-API support * cmake: Address lat9nq's review * cmake: Make VA-API optional * vaapi: Multiple GPU * Apply suggestions from code review Co-authored-by: Ameer J <52414509+ameerj@users.noreply.github.com> * nvdec: Address ameerj's review * codec: Use anonymous instead of static * nvdec: Remove enum and fix memory leak * nvdec: Address ameerj's review * codec: Remove preparation for threading Co-authored-by: Ameer J <52414509+ameerj@users.noreply.github.com>
* Merge pull request #6525 from ameerj/nvdec-fixesFernando S2021-07-151-11/+10
|\ | | | | nvdec: Fix Submit Ioctl data source, vic frame dimension computations
| * vic: Fix dimension compuation of YUV framesameerj2021-07-151-11/+10
| | | | | | | | Fixes out of bound memory crashes in Mario Golf
* | Merge pull request #6537 from Morph1984/warningsbunnei2021-07-062-15/+0
|\ \ | | | | | | general: Enforce multiple warnings in MSVC
| * | video_core: Remove #pragma warning directives for external headersMorph2021-06-282-15/+0
| | |
* | | Slightly refactor NVDEC and codecs for readability and safetyKelebek12021-07-019-354/+521
|/ /
* / codec,vic: Disable warnings in ffmpeg headersReinUsesLisp2021-06-262-4/+29
|/
* vp9: Avoid memcpy with null pointerslat9nq2021-04-051-7/+9
| | | | | | | | Avoid sending null pointer to memcpy as reported by Undefined Behaviour Sanitizer. Replaces the std::memcpy calls in SpliceVectors with std::copy calls. Opting to replace all the memcpy's with copy's. Co-authored-by: LC <mathew1800@gmail.com>
* rebase, fix name shadowing, more constameerj2021-02-131-5/+5
|
* streamline cdma_pusher/command_classesameerj2021-02-131-13/+5
|
* streamline cdma_pusher/command_classesameerj2021-02-133-48/+6
|
* nvdec cleanupameerj2021-02-133-11/+24
|
* video_core: Reimplement the buffer cacheReinUsesLisp2021-02-131-3/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Reimplement the buffer cache using cached bindings and page level granularity for modification tracking. This also drops the usage of shared pointers and virtual functions from the cache. - Bindings are cached, allowing to skip work when the game changes few bits between draws. - OpenGL Assembly shaders no longer copy when a region has been modified from the GPU to emulate constant buffers, instead GL_EXT_memory_object is used to alias sub-buffers within the same allocation. - OpenGL Assembly shaders stream constant buffer data using glProgramBufferParametersIuivNV, from NV_parameter_buffer_object. In theory this should save one hash table resolve inside the driver compared to glBufferSubData. - A new OpenGL stream buffer is implemented based on fences for drivers that are not Nvidia's proprietary, due to their low performance on partial glBufferSubData calls synchronized with 3D rendering (that some games use a lot). - Most optimizations are shared between APIs now, allowing Vulkan to cache more bindings than before, skipping unnecesarry work. This commit adds the necessary infrastructure to use Vulkan object from OpenGL. Overall, it improves performance and fixes some bugs present on the old cache. There are still some edge cases hit by some games that harm performance on some vendors, this are planned to be fixed in later commits.
* common/bit_util: Replace CLZ/CTZ operations with standardized onesLioncash2021-01-151-2/+2
| | | | Makes for less code that we need to maintain.
* remove inaccurate referenceAmeer J2021-01-071-1/+1
| | | Co-authored-by: LC <mathew1800@gmail.com>
* fix for nvdec disabled, cleanup host1xameerj2021-01-072-61/+9
|
* nvdec syncpt incorporationameerj2021-01-072-3/+5
| | | | laying the groundwork for async gpu, although this does not fully implement async nvdec operations
* general: Fix various spelling errorsMorph2021-01-021-1/+1
|
* video_core: Rewrite the texture cacheReinUsesLisp2020-12-301-4/+4
| | | | | | | | | | | | | | The current texture cache has several points that hurt maintainability and performance. It's easy to break unrelated parts of the cache when doing minor changes. The cache can easily forget valuable information about the cached textures by CPU writes or simply by its normal usage.The current texture cache has several points that hurt maintainability and performance. It's easy to break unrelated parts of the cache when doing minor changes. The cache can easily forget valuable information about the cached textures by CPU writes or simply by its normal usage. This commit aims to address those issues.
* video_core: Remove unnecessary enum class casting in logging messagesLioncash2020-12-072-3/+3
| | | | | | | fmt now automatically prints the numeric value of an enum class member by default, so we don't need to use casts any more. Reduces the line noise a bit.
* video_core: Resolve more variable shadowing scenarios pt.2Lioncash2020-12-051-20/+20
| | | | | | | Migrates the video core code closer to enabling variable shadowing warnings as errors. This primarily sorts out shadowing occurrences within the Vulkan code.
* Merge pull request #5124 from lioncash/video-shadowbunnei2020-12-052-2/+2
|\ | | | | video_core: Resolve more variable shadowing scenarios
| * video_core: Resolve more variable shadowing scenariosLioncash2020-12-042-2/+2
| | | | | | | | | | | | Resolves variable shadowing scenarios up to the end of the OpenGL code to make it nicer to review. The rest will be resolved in a following commit.
* | codec: Remove deprecated usage of AVCodecContext::refcounted_framesLioncash2020-12-041-1/+0
|/ | | | | | | | | | | | | | | | This was only necessary for use with the avcodec_decode_video2/avcoded_decode_audio4 APIs which are also deprecated. Given we use avcodec_send_packet/avcodec_receive_frame, this isn't necessary, this is even indicated directly within the FFmpeg API changes document here on 2017-09-26: https://github.com/FFmpeg/FFmpeg/blob/master/doc/APIchanges#L410 This prevents our code from breaking whenever we update to a newer version of FFmpeg in the future if they ever decide to fully remove this API member.
* vp9/vic: Resolve pessimizing movesLioncash2020-12-032-11/+11
| | | | | Removes the usage of moves that don't result in behavior different from a copy, or otherwise would prevent copy elision from occurring.
* Merge pull request #5002 from ameerj/nvdec-frameskipbunnei2020-12-0310-340/+234
|\ | | | | nvdec: Queue and display all decoded frames, cleanup decoders
| * Limit queue size to 10 framesameerj2020-11-261-0/+4
| | | | | | | | Workaround for ZLA, which seems to decode and queue twice as many frames as it displays.
| * Address PR feedbackameerj2020-11-264-32/+33
| | | | | | | | | | | | remove some redundant moves, make deleter match naming guidelines. Co-Authored-By: LC <712067+lioncash@users.noreply.github.com>
| * Queue decoded frames, cleanup decodersameerj2020-11-2510-338/+227
| |
* | codec: Fix `pragma GCC diagnostic pop` missing corresponding pushcomex2020-11-261-0/+1
|/
* nvdec: Make use of [[nodiscard]] where applicableLioncash2020-11-027-16/+16
| | | | | Prevents bugs from occurring where the results of a function are accidentally discarded
* vp9: Be explicit with copy and move operatorsLioncash2020-10-301-0/+18
| | | | | | It's deprecated in the language to autogenerate these if the destructor for a type is specified, so we can explicitly specify how we want these to be generated.
* vp9: Mark functions with [[nodiscard]] where applicableLioncash2020-10-302-13/+13
| | | | | Prevents values from mistakenly being discarded in cases where it's a bug to do so.
* vp9: Provide a default initializer for "hidden" memberLioncash2020-10-301-1/+1
| | | | | | | | | The API of VP9 exposes a WasFrameHidden() function which accesses this member. Given the constructor previously didn't initialize this member, it's a potential vector for an uninitialized read. Instead, we can initialize this to a deterministic value to prevent that from occurring.
* vp9: Make some member functions internally linkedLioncash2020-10-302-58/+54
| | | | | These helper functions don't directly modify any member state and can be hidden from view.
* Merge pull request #4837 from lioncash/nvdec-2bunnei2020-10-2913-88/+81
|\ | | | | nvdec: Minor tidying up
| * h264: Make WriteUe take a u32Lioncash2020-10-272-7/+8
| | | | | | | | Enforces the type of the desired value in calling code.
| * vp9: std::move buffer within ComposeFrameHeader()Lioncash2020-10-271-1/+1
| | | | | | | | We can move the buffer here to avoid a heap reallocation
| * vp9: Remove dead codeLioncash2020-10-271-6/+0
| |
| * vp9: Join declarations with assignmentsLioncash2020-10-271-7/+8
| |
| * vp9: Remove pessimizing movesLioncash2020-10-271-2/+2
| | | | | | | | The move will already occur without std::move.
| * vp9: Resolve variable shadowingLioncash2020-10-271-4/+4
| |
| * nvdec: Tidy up header includesLioncash2020-10-2713-62/+59
| | | | | | | | Prevents a few unnecessary inclusions.
* | Merge pull request #4838 from lioncash/syncmgrbunnei2020-10-292-9/+9
|\ \ | | | | | | sync_manager: Amend parameter order of calls to SyncptIncr constructor
| * | sync_manager: Amend parameter order of calls to SyncptIncr constructorLioncash2020-10-272-9/+9
| |/ | | | | | | Corrects some cases where the arguments would be incorrectly swapped.
* | Merge pull request #4848 from ReinUsesLisp/type-limitsLC2020-10-281-1/+1
|\ \ | | | | | | video_core: Enforce -Werror=type-limits
| * | video_core: Enforce -Werror=type-limitsReinUsesLisp2020-10-281-1/+1
| |/ | | | | | | Silences one warning and avoids introducing more in the future.
* / video_core: Enforce -Wredundant-move and -Wpessimizing-moveReinUsesLisp2020-10-281-1/+1
|/ | | | Silence three warnings and make them errors to avoid introducing more in the future.
* video_core: NVDEC Implementationameerj2020-10-2716-0/+2857
This commit aims to implement the NVDEC (Nvidia Decoder) functionality, with video frame decoding being handled by the FFmpeg library. The process begins with Ioctl commands being sent to the NVDEC and VIC (Video Image Composer) emulated devices. These allocate the necessary GPU buffers for the frame data, along with providing information on the incoming video data. A Submit command then signals the GPU to process and decode the frame data. To decode the frame, the respective codec's header must be manually composed from the information provided by NVDEC, then sent with the raw frame data to the ffmpeg library. Currently, H264 and VP9 are supported, with VP9 having some minor artifacting issues related mainly to the reference frame composition in its uncompressed header. Async GPU is not properly implemented at the moment. Co-Authored-By: David <25727384+ogniK5377@users.noreply.github.com>