From 9ef5c53e52941548c8a7e097ec3df8c07d460567 Mon Sep 17 00:00:00 2001 From: ameerj Date: Mon, 26 Oct 2020 23:07:36 -0400 Subject: video_core: NVDEC Implementation 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> --- src/video_core/command_classes/codecs/vp9.cpp | 1010 +++++++++++++++++++++++++ 1 file changed, 1010 insertions(+) create mode 100644 src/video_core/command_classes/codecs/vp9.cpp (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp new file mode 100644 index 0000000000..3bae0bb5d9 --- /dev/null +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -0,0 +1,1010 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include // for std::memcpy +#include +#include "video_core/command_classes/codecs/vp9.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +namespace Tegra::Decoder { + +// Default compressed header probabilities once frame context resets +constexpr Vp9EntropyProbs default_probs{ + .y_mode_prob{ + 65, 32, 18, 144, 162, 194, 41, 51, 98, 132, 68, 18, 165, 217, 196, 45, 40, 78, + 173, 80, 19, 176, 240, 193, 64, 35, 46, 221, 135, 38, 194, 248, 121, 96, 85, 29, + }, + .partition_prob{ + 199, 122, 141, 0, 147, 63, 159, 0, 148, 133, 118, 0, 121, 104, 114, 0, + 174, 73, 87, 0, 92, 41, 83, 0, 82, 99, 50, 0, 53, 39, 39, 0, + 177, 58, 59, 0, 68, 26, 63, 0, 52, 79, 25, 0, 17, 14, 12, 0, + 222, 34, 30, 0, 72, 16, 44, 0, 58, 32, 12, 0, 10, 7, 6, 0, + }, + .coef_probs{ + 195, 29, 183, 0, 84, 49, 136, 0, 8, 42, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 31, 107, 169, 0, 35, 99, 159, 0, 17, 82, 140, 0, 8, 66, 114, 0, + 2, 44, 76, 0, 1, 19, 32, 0, 40, 132, 201, 0, 29, 114, 187, 0, 13, 91, 157, 0, + 7, 75, 127, 0, 3, 58, 95, 0, 1, 28, 47, 0, 69, 142, 221, 0, 42, 122, 201, 0, + 15, 91, 159, 0, 6, 67, 121, 0, 1, 42, 77, 0, 1, 17, 31, 0, 102, 148, 228, 0, + 67, 117, 204, 0, 17, 82, 154, 0, 6, 59, 114, 0, 2, 39, 75, 0, 1, 15, 29, 0, + 156, 57, 233, 0, 119, 57, 212, 0, 58, 48, 163, 0, 29, 40, 124, 0, 12, 30, 81, 0, + 3, 12, 31, 0, 191, 107, 226, 0, 124, 117, 204, 0, 25, 99, 155, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 29, 148, 210, 0, 37, 126, 194, 0, 8, 93, 157, 0, + 2, 68, 118, 0, 1, 39, 69, 0, 1, 17, 33, 0, 41, 151, 213, 0, 27, 123, 193, 0, + 3, 82, 144, 0, 1, 58, 105, 0, 1, 32, 60, 0, 1, 13, 26, 0, 59, 159, 220, 0, + 23, 126, 198, 0, 4, 88, 151, 0, 1, 66, 114, 0, 1, 38, 71, 0, 1, 18, 34, 0, + 114, 136, 232, 0, 51, 114, 207, 0, 11, 83, 155, 0, 3, 56, 105, 0, 1, 33, 65, 0, + 1, 17, 34, 0, 149, 65, 234, 0, 121, 57, 215, 0, 61, 49, 166, 0, 28, 36, 114, 0, + 12, 25, 76, 0, 3, 16, 42, 0, 214, 49, 220, 0, 132, 63, 188, 0, 42, 65, 137, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 137, 221, 0, 104, 131, 216, 0, + 49, 111, 192, 0, 21, 87, 155, 0, 2, 49, 87, 0, 1, 16, 28, 0, 89, 163, 230, 0, + 90, 137, 220, 0, 29, 100, 183, 0, 10, 70, 135, 0, 2, 42, 81, 0, 1, 17, 33, 0, + 108, 167, 237, 0, 55, 133, 222, 0, 15, 97, 179, 0, 4, 72, 135, 0, 1, 45, 85, 0, + 1, 19, 38, 0, 124, 146, 240, 0, 66, 124, 224, 0, 17, 88, 175, 0, 4, 58, 122, 0, + 1, 36, 75, 0, 1, 18, 37, 0, 141, 79, 241, 0, 126, 70, 227, 0, 66, 58, 182, 0, + 30, 44, 136, 0, 12, 34, 96, 0, 2, 20, 47, 0, 229, 99, 249, 0, 143, 111, 235, 0, + 46, 109, 192, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 82, 158, 236, 0, + 94, 146, 224, 0, 25, 117, 191, 0, 9, 87, 149, 0, 3, 56, 99, 0, 1, 33, 57, 0, + 83, 167, 237, 0, 68, 145, 222, 0, 10, 103, 177, 0, 2, 72, 131, 0, 1, 41, 79, 0, + 1, 20, 39, 0, 99, 167, 239, 0, 47, 141, 224, 0, 10, 104, 178, 0, 2, 73, 133, 0, + 1, 44, 85, 0, 1, 22, 47, 0, 127, 145, 243, 0, 71, 129, 228, 0, 17, 93, 177, 0, + 3, 61, 124, 0, 1, 41, 84, 0, 1, 21, 52, 0, 157, 78, 244, 0, 140, 72, 231, 0, + 69, 58, 184, 0, 31, 44, 137, 0, 14, 38, 105, 0, 8, 23, 61, 0, 125, 34, 187, 0, + 52, 41, 133, 0, 6, 31, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 37, 109, 153, 0, 51, 102, 147, 0, 23, 87, 128, 0, 8, 67, 101, 0, 1, 41, 63, 0, + 1, 19, 29, 0, 31, 154, 185, 0, 17, 127, 175, 0, 6, 96, 145, 0, 2, 73, 114, 0, + 1, 51, 82, 0, 1, 28, 45, 0, 23, 163, 200, 0, 10, 131, 185, 0, 2, 93, 148, 0, + 1, 67, 111, 0, 1, 41, 69, 0, 1, 14, 24, 0, 29, 176, 217, 0, 12, 145, 201, 0, + 3, 101, 156, 0, 1, 69, 111, 0, 1, 39, 63, 0, 1, 14, 23, 0, 57, 192, 233, 0, + 25, 154, 215, 0, 6, 109, 167, 0, 3, 78, 118, 0, 1, 48, 69, 0, 1, 21, 29, 0, + 202, 105, 245, 0, 108, 106, 216, 0, 18, 90, 144, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 33, 172, 219, 0, 64, 149, 206, 0, 14, 117, 177, 0, 5, 90, 141, 0, + 2, 61, 95, 0, 1, 37, 57, 0, 33, 179, 220, 0, 11, 140, 198, 0, 1, 89, 148, 0, + 1, 60, 104, 0, 1, 33, 57, 0, 1, 12, 21, 0, 30, 181, 221, 0, 8, 141, 198, 0, + 1, 87, 145, 0, 1, 58, 100, 0, 1, 31, 55, 0, 1, 12, 20, 0, 32, 186, 224, 0, + 7, 142, 198, 0, 1, 86, 143, 0, 1, 58, 100, 0, 1, 31, 55, 0, 1, 12, 22, 0, + 57, 192, 227, 0, 20, 143, 204, 0, 3, 96, 154, 0, 1, 68, 112, 0, 1, 42, 69, 0, + 1, 19, 32, 0, 212, 35, 215, 0, 113, 47, 169, 0, 29, 48, 105, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 74, 129, 203, 0, 106, 120, 203, 0, 49, 107, 178, 0, + 19, 84, 144, 0, 4, 50, 84, 0, 1, 15, 25, 0, 71, 172, 217, 0, 44, 141, 209, 0, + 15, 102, 173, 0, 6, 76, 133, 0, 2, 51, 89, 0, 1, 24, 42, 0, 64, 185, 231, 0, + 31, 148, 216, 0, 8, 103, 175, 0, 3, 74, 131, 0, 1, 46, 81, 0, 1, 18, 30, 0, + 65, 196, 235, 0, 25, 157, 221, 0, 5, 105, 174, 0, 1, 67, 120, 0, 1, 38, 69, 0, + 1, 15, 30, 0, 65, 204, 238, 0, 30, 156, 224, 0, 7, 107, 177, 0, 2, 70, 124, 0, + 1, 42, 73, 0, 1, 18, 34, 0, 225, 86, 251, 0, 144, 104, 235, 0, 42, 99, 181, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 175, 239, 0, 112, 165, 229, 0, + 29, 136, 200, 0, 12, 103, 162, 0, 6, 77, 123, 0, 2, 53, 84, 0, 75, 183, 239, 0, + 30, 155, 221, 0, 3, 106, 171, 0, 1, 74, 128, 0, 1, 44, 76, 0, 1, 17, 28, 0, + 73, 185, 240, 0, 27, 159, 222, 0, 2, 107, 172, 0, 1, 75, 127, 0, 1, 42, 73, 0, + 1, 17, 29, 0, 62, 190, 238, 0, 21, 159, 222, 0, 2, 107, 172, 0, 1, 72, 122, 0, + 1, 40, 71, 0, 1, 18, 32, 0, 61, 199, 240, 0, 27, 161, 226, 0, 4, 113, 180, 0, + 1, 76, 129, 0, 1, 46, 80, 0, 1, 23, 41, 0, 7, 27, 153, 0, 5, 30, 95, 0, + 1, 16, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 50, 75, 127, 0, + 57, 75, 124, 0, 27, 67, 108, 0, 10, 54, 86, 0, 1, 33, 52, 0, 1, 12, 18, 0, + 43, 125, 151, 0, 26, 108, 148, 0, 7, 83, 122, 0, 2, 59, 89, 0, 1, 38, 60, 0, + 1, 17, 27, 0, 23, 144, 163, 0, 13, 112, 154, 0, 2, 75, 117, 0, 1, 50, 81, 0, + 1, 31, 51, 0, 1, 14, 23, 0, 18, 162, 185, 0, 6, 123, 171, 0, 1, 78, 125, 0, + 1, 51, 86, 0, 1, 31, 54, 0, 1, 14, 23, 0, 15, 199, 227, 0, 3, 150, 204, 0, + 1, 91, 146, 0, 1, 55, 95, 0, 1, 30, 53, 0, 1, 11, 20, 0, 19, 55, 240, 0, + 19, 59, 196, 0, 3, 52, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 41, 166, 207, 0, 104, 153, 199, 0, 31, 123, 181, 0, 14, 101, 152, 0, 5, 72, 106, 0, + 1, 36, 52, 0, 35, 176, 211, 0, 12, 131, 190, 0, 2, 88, 144, 0, 1, 60, 101, 0, + 1, 36, 60, 0, 1, 16, 28, 0, 28, 183, 213, 0, 8, 134, 191, 0, 1, 86, 142, 0, + 1, 56, 96, 0, 1, 30, 53, 0, 1, 12, 20, 0, 20, 190, 215, 0, 4, 135, 192, 0, + 1, 84, 139, 0, 1, 53, 91, 0, 1, 28, 49, 0, 1, 11, 20, 0, 13, 196, 216, 0, + 2, 137, 192, 0, 1, 86, 143, 0, 1, 57, 99, 0, 1, 32, 56, 0, 1, 13, 24, 0, + 211, 29, 217, 0, 96, 47, 156, 0, 22, 43, 87, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 78, 120, 193, 0, 111, 116, 186, 0, 46, 102, 164, 0, 15, 80, 128, 0, + 2, 49, 76, 0, 1, 18, 28, 0, 71, 161, 203, 0, 42, 132, 192, 0, 10, 98, 150, 0, + 3, 69, 109, 0, 1, 44, 70, 0, 1, 18, 29, 0, 57, 186, 211, 0, 30, 140, 196, 0, + 4, 93, 146, 0, 1, 62, 102, 0, 1, 38, 65, 0, 1, 16, 27, 0, 47, 199, 217, 0, + 14, 145, 196, 0, 1, 88, 142, 0, 1, 57, 98, 0, 1, 36, 62, 0, 1, 15, 26, 0, + 26, 219, 229, 0, 5, 155, 207, 0, 1, 94, 151, 0, 1, 60, 104, 0, 1, 36, 62, 0, + 1, 16, 28, 0, 233, 29, 248, 0, 146, 47, 220, 0, 43, 52, 140, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 100, 163, 232, 0, 179, 161, 222, 0, 63, 142, 204, 0, + 37, 113, 174, 0, 26, 89, 137, 0, 18, 68, 97, 0, 85, 181, 230, 0, 32, 146, 209, 0, + 7, 100, 164, 0, 3, 71, 121, 0, 1, 45, 77, 0, 1, 18, 30, 0, 65, 187, 230, 0, + 20, 148, 207, 0, 2, 97, 159, 0, 1, 68, 116, 0, 1, 40, 70, 0, 1, 14, 29, 0, + 40, 194, 227, 0, 8, 147, 204, 0, 1, 94, 155, 0, 1, 65, 112, 0, 1, 39, 66, 0, + 1, 14, 26, 0, 16, 208, 228, 0, 3, 151, 207, 0, 1, 98, 160, 0, 1, 67, 117, 0, + 1, 41, 74, 0, 1, 17, 31, 0, 17, 38, 140, 0, 7, 34, 80, 0, 1, 17, 29, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 37, 75, 128, 0, 41, 76, 128, 0, + 26, 66, 116, 0, 12, 52, 94, 0, 2, 32, 55, 0, 1, 10, 16, 0, 50, 127, 154, 0, + 37, 109, 152, 0, 16, 82, 121, 0, 5, 59, 85, 0, 1, 35, 54, 0, 1, 13, 20, 0, + 40, 142, 167, 0, 17, 110, 157, 0, 2, 71, 112, 0, 1, 44, 72, 0, 1, 27, 45, 0, + 1, 11, 17, 0, 30, 175, 188, 0, 9, 124, 169, 0, 1, 74, 116, 0, 1, 48, 78, 0, + 1, 30, 49, 0, 1, 11, 18, 0, 10, 222, 223, 0, 2, 150, 194, 0, 1, 83, 128, 0, + 1, 48, 79, 0, 1, 27, 45, 0, 1, 11, 17, 0, 36, 41, 235, 0, 29, 36, 193, 0, + 10, 27, 111, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 165, 222, 0, + 177, 162, 215, 0, 110, 135, 195, 0, 57, 113, 168, 0, 23, 83, 120, 0, 10, 49, 61, 0, + 85, 190, 223, 0, 36, 139, 200, 0, 5, 90, 146, 0, 1, 60, 103, 0, 1, 38, 65, 0, + 1, 18, 30, 0, 72, 202, 223, 0, 23, 141, 199, 0, 2, 86, 140, 0, 1, 56, 97, 0, + 1, 36, 61, 0, 1, 16, 27, 0, 55, 218, 225, 0, 13, 145, 200, 0, 1, 86, 141, 0, + 1, 57, 99, 0, 1, 35, 61, 0, 1, 13, 22, 0, 15, 235, 212, 0, 1, 132, 184, 0, + 1, 84, 139, 0, 1, 57, 97, 0, 1, 34, 56, 0, 1, 14, 23, 0, 181, 21, 201, 0, + 61, 37, 123, 0, 10, 38, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 47, 106, 172, 0, 95, 104, 173, 0, 42, 93, 159, 0, 18, 77, 131, 0, 4, 50, 81, 0, + 1, 17, 23, 0, 62, 147, 199, 0, 44, 130, 189, 0, 28, 102, 154, 0, 18, 75, 115, 0, + 2, 44, 65, 0, 1, 12, 19, 0, 55, 153, 210, 0, 24, 130, 194, 0, 3, 93, 146, 0, + 1, 61, 97, 0, 1, 31, 50, 0, 1, 10, 16, 0, 49, 186, 223, 0, 17, 148, 204, 0, + 1, 96, 142, 0, 1, 53, 83, 0, 1, 26, 44, 0, 1, 11, 17, 0, 13, 217, 212, 0, + 2, 136, 180, 0, 1, 78, 124, 0, 1, 50, 83, 0, 1, 29, 49, 0, 1, 14, 23, 0, + 197, 13, 247, 0, 82, 17, 222, 0, 25, 17, 162, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 126, 186, 247, 0, 234, 191, 243, 0, 176, 177, 234, 0, 104, 158, 220, 0, + 66, 128, 186, 0, 55, 90, 137, 0, 111, 197, 242, 0, 46, 158, 219, 0, 9, 104, 171, 0, + 2, 65, 125, 0, 1, 44, 80, 0, 1, 17, 91, 0, 104, 208, 245, 0, 39, 168, 224, 0, + 3, 109, 162, 0, 1, 79, 124, 0, 1, 50, 102, 0, 1, 43, 102, 0, 84, 220, 246, 0, + 31, 177, 231, 0, 2, 115, 180, 0, 1, 79, 134, 0, 1, 55, 77, 0, 1, 60, 79, 0, + 43, 243, 240, 0, 8, 180, 217, 0, 1, 115, 166, 0, 1, 84, 121, 0, 1, 51, 67, 0, + 1, 16, 6, 0, + }, + .switchable_interp_prob{235, 162, 36, 255, 34, 3, 149, 144}, + .inter_mode_prob{ + 2, 173, 34, 0, 7, 145, 85, 0, 7, 166, 63, 0, 7, 94, + 66, 0, 8, 64, 46, 0, 17, 81, 31, 0, 25, 29, 30, 0, + }, + .intra_inter_prob{9, 102, 187, 225}, + .comp_inter_prob{9, 102, 187, 225, 0}, + .single_ref_prob{33, 16, 77, 74, 142, 142, 172, 170, 238, 247}, + .comp_ref_prob{50, 126, 123, 221, 226}, + .tx_32x32_prob{3, 136, 37, 5, 52, 13}, + .tx_16x16_prob{20, 152, 15, 101}, + .tx_8x8_prob{100, 66}, + .skip_probs{192, 128, 64}, + .joints{32, 64, 96}, + .sign{128, 128}, + .classes{ + 224, 144, 192, 168, 192, 176, 192, 198, 198, 245, + 216, 128, 176, 160, 176, 176, 192, 198, 198, 208, + }, + .class_0{216, 208}, + .prob_bits{ + 136, 140, 148, 160, 176, 192, 224, 234, 234, 240, + 136, 140, 148, 160, 176, 192, 224, 234, 234, 240, + }, + .class_0_fr{128, 128, 64, 96, 112, 64, 128, 128, 64, 96, 112, 64}, + .fr{64, 96, 64, 64, 96, 64}, + .class_0_hp{160, 160}, + .high_precision{128, 128}, +}; + +VP9::VP9(GPU& gpu) : gpu(gpu) {} + +VP9::~VP9() = default; + +void VP9::WriteProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const bool update = new_prob != old_prob; + + writer.Write(update, diff_update_probability); + + if (update) { + WriteProbabilityDelta(writer, new_prob, old_prob); + } +} +template +void VP9::WriteProbabilityUpdate(VpxRangeEncoder& writer, const std::array& new_prob, + const std::array& old_prob) { + for (std::size_t offset = 0; offset < new_prob.size(); ++offset) { + WriteProbabilityUpdate(writer, new_prob[offset], old_prob[offset]); + } +} + +template +void VP9::WriteProbabilityUpdateAligned4(VpxRangeEncoder& writer, const std::array& new_prob, + const std::array& old_prob) { + for (std::size_t offset = 0; offset < new_prob.size(); offset += 4) { + WriteProbabilityUpdate(writer, new_prob[offset + 0], old_prob[offset + 0]); + WriteProbabilityUpdate(writer, new_prob[offset + 1], old_prob[offset + 1]); + WriteProbabilityUpdate(writer, new_prob[offset + 2], old_prob[offset + 2]); + } +} + +void VP9::WriteProbabilityDelta(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const int delta = RemapProbability(new_prob, old_prob); + + EncodeTermSubExp(writer, delta); +} + +s32 VP9::RemapProbability(s32 new_prob, s32 old_prob) { + new_prob--; + old_prob--; + + std::size_t index{}; + + if (old_prob * 2 <= 0xff) { + index = static_cast(std::max(0, RecenterNonNeg(new_prob, old_prob) - 1)); + } else { + index = static_cast( + std::max(0, RecenterNonNeg(0xff - 1 - new_prob, 0xff - 1 - old_prob) - 1)); + } + + return map_lut[index]; +} + +s32 VP9::RecenterNonNeg(s32 new_prob, s32 old_prob) { + if (new_prob > old_prob * 2) { + return new_prob; + } else if (new_prob >= old_prob) { + return (new_prob - old_prob) * 2; + } else { + return (old_prob - new_prob) * 2 - 1; + } +} + +void VP9::EncodeTermSubExp(VpxRangeEncoder& writer, s32 value) { + if (WriteLessThan(writer, value, 16)) { + writer.Write(value, 4); + } else if (WriteLessThan(writer, value, 32)) { + writer.Write(value - 16, 4); + } else if (WriteLessThan(writer, value, 64)) { + writer.Write(value - 32, 5); + } else { + value -= 64; + + constexpr s32 size = 8; + + const s32 mask = (1 << size) - 191; + + const s32 delta = value - mask; + + if (delta < 0) { + writer.Write(value, size - 1); + } else { + writer.Write(delta / 2 + mask, size - 1); + writer.Write(delta & 1, 1); + } + } +} + +bool VP9::WriteLessThan(VpxRangeEncoder& writer, s32 value, s32 test) { + const bool is_lt = value < test; + writer.Write(!is_lt); + return is_lt; +} + +void VP9::WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, + const std::array& new_prob, + const std::array& old_prob) { + // Note: There's 1 byte added on each packet for alignment, + // this byte is ignored when doing updates. + constexpr s32 block_bytes = 2 * 2 * 6 * 6 * 4; + + const auto needs_update = [&](s32 base_index) -> bool { + s32 index = base_index; + for (s32 i = 0; i < 2; i++) { + for (s32 j = 0; j < 2; j++) { + for (s32 k = 0; k < 6; k++) { + for (s32 l = 0; l < 6; l++) { + if (new_prob[index + 0] != old_prob[index + 0] || + new_prob[index + 1] != old_prob[index + 1] || + new_prob[index + 2] != old_prob[index + 2]) { + return true; + } + + index += 4; + } + } + } + } + return false; + }; + + for (s32 block_index = 0; block_index < 4; block_index++) { + const s32 base_index = block_index * block_bytes; + const bool update = needs_update(base_index); + writer.Write(update); + + if (update) { + s32 index = base_index; + for (s32 i = 0; i < 2; i++) { + for (s32 j = 0; j < 2; j++) { + for (s32 k = 0; k < 6; k++) { + for (s32 l = 0; l < 6; l++) { + if (k != 0 || l < 3) { + WriteProbabilityUpdate(writer, new_prob[index + 0], + old_prob[index + 0]); + WriteProbabilityUpdate(writer, new_prob[index + 1], + old_prob[index + 1]); + WriteProbabilityUpdate(writer, new_prob[index + 2], + old_prob[index + 2]); + } + index += 4; + } + } + } + } + } + + if (block_index == tx_mode) { + break; + } + } +} + +void VP9::WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const bool update = new_prob != old_prob; + writer.Write(update, diff_update_probability); + + if (update) { + writer.Write(new_prob >> 1, 7); + } +} + +s32 VP9::CalcMinLog2TileCols(s32 frame_width) { + const s32 sb64_cols = (frame_width + 63) / 64; + s32 min_log2 = 0; + + while ((64 << min_log2) < sb64_cols) { + min_log2++; + } + + return min_log2; +} + +s32 VP9::CalcMaxLog2TileCols(s32 frameWidth) { + const s32 sb64_cols = (frameWidth + 63) / 64; + s32 max_log2 = 1; + + while ((sb64_cols >> max_log2) >= 4) { + max_log2++; + } + + return max_log2 - 1; +} + +Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) { + PictureInfo picture_info{}; + gpu.MemoryManager().ReadBlock(state.picture_info_offset, &picture_info, sizeof(PictureInfo)); + Vp9PictureInfo vp9_info = picture_info.Convert(); + + InsertEntropy(state.vp9_entropy_probs_offset, vp9_info.entropy); + + // surface_luma_offset[0:3] contains the address of the reference frame offsets in the following + // order: last, golden, altref, current. It may be worthwhile to track the updates done here + // to avoid buffering frame data needed for reference frame updating in the header composition. + std::memcpy(vp9_info.frame_offsets.data(), state.surface_luma_offset.data(), 4 * sizeof(u64)); + + return std::move(vp9_info); +} + +void VP9::InsertEntropy(u64 offset, Vp9EntropyProbs& dst) { + EntropyProbs entropy{}; + gpu.MemoryManager().ReadBlock(offset, &entropy, sizeof(EntropyProbs)); + entropy.Convert(dst); +} + +Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) { + Vp9FrameContainer frame{}; + { + gpu.SyncGuestHost(); + frame.info = std::move(GetVp9PictureInfo(state)); + + frame.bit_stream.resize(frame.info.bitstream_size); + gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, frame.bit_stream.data(), + frame.info.bitstream_size); + } + // Buffer two frames, saving the last show frame info + if (next_next_frame.bit_stream.size() != 0) { + Vp9FrameContainer temp{ + .info = frame.info, + .bit_stream = frame.bit_stream, + }; + next_next_frame.info.show_frame = frame.info.last_frame_shown; + frame.info = next_next_frame.info; + frame.bit_stream = next_next_frame.bit_stream; + next_next_frame = std::move(temp); + + if (next_frame.bit_stream.size() != 0) { + Vp9FrameContainer temp{ + .info = frame.info, + .bit_stream = frame.bit_stream, + }; + next_frame.info.show_frame = frame.info.last_frame_shown; + frame.info = next_frame.info; + frame.bit_stream = next_frame.bit_stream; + next_frame = std::move(temp); + } else { + next_frame.info = frame.info; + next_frame.bit_stream = frame.bit_stream; + } + } else { + next_next_frame.info = frame.info; + next_next_frame.bit_stream = frame.bit_stream; + } + return frame; +} + +std::vector VP9::ComposeCompressedHeader() { + VpxRangeEncoder writer{}; + + if (!current_frame_info.lossless) { + if (static_cast(current_frame_info.transform_mode) >= 3) { + writer.Write(3, 2); + writer.Write(current_frame_info.transform_mode == 4); + } else { + writer.Write(current_frame_info.transform_mode, 2); + } + } + + if (current_frame_info.transform_mode == 4) { + // tx_mode_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_8x8_prob, + prev_frame_probs.tx_8x8_prob); + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_16x16_prob, + prev_frame_probs.tx_16x16_prob); + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_32x32_prob, + prev_frame_probs.tx_32x32_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.tx_8x8_prob = current_frame_info.entropy.tx_8x8_prob; + prev_frame_probs.tx_16x16_prob = current_frame_info.entropy.tx_16x16_prob; + prev_frame_probs.tx_32x32_prob = current_frame_info.entropy.tx_32x32_prob; + } + } + // read_coef_probs() in the spec + WriteCoefProbabilityUpdate(writer, current_frame_info.transform_mode, + current_frame_info.entropy.coef_probs, prev_frame_probs.coef_probs); + // read_skip_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.skip_probs, + prev_frame_probs.skip_probs); + + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.coef_probs = current_frame_info.entropy.coef_probs; + prev_frame_probs.skip_probs = current_frame_info.entropy.skip_probs; + } + + if (!current_frame_info.intra_only) { + // read_inter_probs() in the spec + WriteProbabilityUpdateAligned4(writer, current_frame_info.entropy.inter_mode_prob, + prev_frame_probs.inter_mode_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.inter_mode_prob = current_frame_info.entropy.inter_mode_prob; + } + + if (current_frame_info.interp_filter == 4) { + // read_interp_filter_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.switchable_interp_prob, + prev_frame_probs.switchable_interp_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.switchable_interp_prob = + current_frame_info.entropy.switchable_interp_prob; + } + } + + // read_is_inter_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.intra_inter_prob, + prev_frame_probs.intra_inter_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.intra_inter_prob = current_frame_info.entropy.intra_inter_prob; + } + // frame_reference_mode() in the spec + if ((current_frame_info.ref_frame_sign_bias[1] & 1) != + (current_frame_info.ref_frame_sign_bias[2] & 1) || + (current_frame_info.ref_frame_sign_bias[1] & 1) != + (current_frame_info.ref_frame_sign_bias[3] & 1)) { + if (current_frame_info.reference_mode >= 1) { + writer.Write(1, 1); + writer.Write(current_frame_info.reference_mode == 2); + } else { + writer.Write(0, 1); + } + } + + // frame_reference_mode_probs() in the spec + if (current_frame_info.reference_mode == 2) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.comp_inter_prob, + prev_frame_probs.comp_inter_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.comp_inter_prob = current_frame_info.entropy.comp_inter_prob; + } + } + + if (current_frame_info.reference_mode != 1) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.single_ref_prob, + prev_frame_probs.single_ref_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.single_ref_prob = current_frame_info.entropy.single_ref_prob; + } + } + + if (current_frame_info.reference_mode != 0) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.comp_ref_prob, + prev_frame_probs.comp_ref_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.comp_ref_prob = current_frame_info.entropy.comp_ref_prob; + } + } + + // read_y_mode_probs + for (std::size_t index = 0; index < current_frame_info.entropy.y_mode_prob.size(); + ++index) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.y_mode_prob[index], + prev_frame_probs.y_mode_prob[index]); + } + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.y_mode_prob = current_frame_info.entropy.y_mode_prob; + } + // read_partition_probs + WriteProbabilityUpdateAligned4(writer, current_frame_info.entropy.partition_prob, + prev_frame_probs.partition_prob); + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.partition_prob = current_frame_info.entropy.partition_prob; + } + + // mv_probs + for (s32 i = 0; i < 3; i++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.joints[i], + prev_frame_probs.joints[i]); + } + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.joints = current_frame_info.entropy.joints; + } + + for (s32 i = 0; i < 2; i++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.sign[i], + prev_frame_probs.sign[i]); + + for (s32 j = 0; j < 10; j++) { + const int index = i * 10 + j; + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.classes[index], + prev_frame_probs.classes[index]); + } + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0[i], + prev_frame_probs.class_0[i]); + + for (s32 j = 0; j < 10; j++) { + const int index = i * 10 + j; + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.prob_bits[index], + prev_frame_probs.prob_bits[index]); + } + } + + for (s32 i = 0; i < 2; i++) { + for (s32 j = 0; j < 2; j++) { + for (s32 k = 0; k < 3; k++) { + const int index = i * 2 * 3 + j * 3 + k; + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0_fr[index], + prev_frame_probs.class_0_fr[index]); + } + } + + for (s32 j = 0; j < 3; j++) { + const int index = i * 3 + j; + + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.fr[index], + prev_frame_probs.fr[index]); + } + } + + if (current_frame_info.allow_high_precision_mv) { + for (s32 index = 0; index < 2; index++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0_hp[index], + prev_frame_probs.class_0_hp[index]); + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.high_precision[index], + prev_frame_probs.high_precision[index]); + } + } + + // save previous probs + if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + prev_frame_probs.sign = current_frame_info.entropy.sign; + prev_frame_probs.classes = current_frame_info.entropy.classes; + prev_frame_probs.class_0 = current_frame_info.entropy.class_0; + prev_frame_probs.prob_bits = current_frame_info.entropy.prob_bits; + prev_frame_probs.class_0_fr = current_frame_info.entropy.class_0_fr; + prev_frame_probs.fr = current_frame_info.entropy.fr; + prev_frame_probs.class_0_hp = current_frame_info.entropy.class_0_hp; + prev_frame_probs.high_precision = current_frame_info.entropy.high_precision; + } + } + + writer.End(); + return writer.GetBuffer(); + + const auto writer_bytearray = writer.GetBuffer(); + + std::vector compressed_header(writer_bytearray.size()); + std::memcpy(compressed_header.data(), writer_bytearray.data(), writer_bytearray.size()); + return compressed_header; +} + +VpxBitStreamWriter VP9::ComposeUncompressedHeader() { + VpxBitStreamWriter uncomp_writer{}; + + uncomp_writer.WriteU(2, 2); // Frame marker. + uncomp_writer.WriteU(0, 2); // Profile. + uncomp_writer.WriteBit(false); // Show existing frame. + uncomp_writer.WriteBit(!current_frame_info.is_key_frame); // is key frame? + uncomp_writer.WriteBit(current_frame_info.show_frame); // show frame? + uncomp_writer.WriteBit(current_frame_info.error_resilient_mode); // error reslience + + if (current_frame_info.is_key_frame) { + uncomp_writer.WriteU(frame_sync_code, 24); + uncomp_writer.WriteU(0, 3); // Color space. + uncomp_writer.WriteU(0, 1); // Color range. + uncomp_writer.WriteU(current_frame_info.frame_size.width - 1, 16); + uncomp_writer.WriteU(current_frame_info.frame_size.height - 1, 16); + uncomp_writer.WriteBit(false); // Render and frame size different. + + // Reset context + prev_frame_probs = default_probs; + swap_next_golden = false; + loop_filter_ref_deltas.fill(0); + loop_filter_mode_deltas.fill(0); + + // allow frames offsets to stabilize before checking for golden frames + grace_period = 4; + + // On key frames, all frame slots are set to the current frame, + // so the value of the selected slot doesn't really matter. + frame_ctxs.fill({current_frame_number, false, default_probs}); + + // intra only, meaning the frame can be recreated with no other references + current_frame_info.intra_only = true; + + } else { + std::array ref_frame_index; + + if (!current_frame_info.show_frame) { + uncomp_writer.WriteBit(current_frame_info.intra_only); + if (!current_frame_info.last_frame_was_key) { + swap_next_golden = !swap_next_golden; + } + } else { + current_frame_info.intra_only = false; + } + if (!current_frame_info.error_resilient_mode) { + uncomp_writer.WriteU(0, 2); // Reset frame context. + } + + // Last, Golden, Altref frames + ref_frame_index = std::array{0, 1, 2}; + + // set when next frame is hidden + // altref and golden references are swapped + if (swap_next_golden) { + ref_frame_index = std::array{0, 2, 1}; + } + + // update Last Frame + u64 refresh_frame_flags = 1; + + // golden frame may refresh, determined if the next golden frame offset is changed + bool golden_refresh = false; + if (grace_period <= 0) { + for (s32 index = 1; index < 3; ++index) { + if (current_frame_info.frame_offsets[index] != + next_frame.info.frame_offsets[index]) { + current_frame_info.refresh_frame[index] = true; + golden_refresh = true; + grace_period = 3; + } + } + } + + if (current_frame_info.show_frame && + (!next_frame.info.show_frame || next_frame.info.is_key_frame)) { + // Update golden frame + refresh_frame_flags = swap_next_golden ? 2 : 4; + } + + if (!current_frame_info.show_frame) { + // Update altref + refresh_frame_flags = swap_next_golden ? 2 : 4; + } else if (golden_refresh) { + refresh_frame_flags = 3; + } + + if (current_frame_info.intra_only) { + uncomp_writer.WriteU(frame_sync_code, 24); + uncomp_writer.WriteU(static_cast(refresh_frame_flags), 8); + uncomp_writer.WriteU(current_frame_info.frame_size.width - 1, 16); + uncomp_writer.WriteU(current_frame_info.frame_size.height - 1, 16); + uncomp_writer.WriteBit(false); // Render and frame size different. + } else { + uncomp_writer.WriteU(static_cast(refresh_frame_flags), 8); + + for (s32 index = 1; index < 4; index++) { + uncomp_writer.WriteU(ref_frame_index[index - 1], 3); + uncomp_writer.WriteU(current_frame_info.ref_frame_sign_bias[index], 1); + } + + uncomp_writer.WriteBit(true); // Frame size with refs. + uncomp_writer.WriteBit(false); // Render and frame size different. + uncomp_writer.WriteBit(current_frame_info.allow_high_precision_mv); + uncomp_writer.WriteBit(current_frame_info.interp_filter == 4); + + if (current_frame_info.interp_filter != 4) { + uncomp_writer.WriteU(current_frame_info.interp_filter, 2); + } + } + } + + if (!current_frame_info.error_resilient_mode) { + uncomp_writer.WriteBit(true); // Refresh frame context. where do i get this info from? + uncomp_writer.WriteBit(true); // Frame parallel decoding mode. + } + + int frame_ctx_idx = 0; + if (!current_frame_info.show_frame) { + frame_ctx_idx = 1; + } + + uncomp_writer.WriteU(frame_ctx_idx, 2); // Frame context index. + prev_frame_probs = + frame_ctxs[frame_ctx_idx].probs; // reference probabilities for compressed header + frame_ctxs[frame_ctx_idx] = {current_frame_number, false, current_frame_info.entropy}; + + uncomp_writer.WriteU(current_frame_info.first_level, 6); + uncomp_writer.WriteU(current_frame_info.sharpness_level, 3); + uncomp_writer.WriteBit(current_frame_info.mode_ref_delta_enabled); + + if (current_frame_info.mode_ref_delta_enabled) { + // check if ref deltas are different, update accordingly + std::array update_loop_filter_ref_deltas; + std::array update_loop_filter_mode_deltas; + + bool loop_filter_delta_update = false; + + for (std::size_t index = 0; index < current_frame_info.ref_deltas.size(); index++) { + const s8 old_deltas = loop_filter_ref_deltas[index]; + const s8 new_deltas = current_frame_info.ref_deltas[index]; + + loop_filter_delta_update |= + (update_loop_filter_ref_deltas[index] = old_deltas != new_deltas); + } + + for (std::size_t index = 0; index < current_frame_info.mode_deltas.size(); index++) { + const s8 old_deltas = loop_filter_mode_deltas[index]; + const s8 new_deltas = current_frame_info.mode_deltas[index]; + + loop_filter_delta_update |= + (update_loop_filter_mode_deltas[index] = old_deltas != new_deltas); + } + + uncomp_writer.WriteBit(loop_filter_delta_update); + + if (loop_filter_delta_update) { + for (std::size_t index = 0; index < current_frame_info.ref_deltas.size(); index++) { + uncomp_writer.WriteBit(update_loop_filter_ref_deltas[index]); + + if (update_loop_filter_ref_deltas[index]) { + uncomp_writer.WriteS(current_frame_info.ref_deltas[index], 6); + } + } + + for (std::size_t index = 0; index < current_frame_info.mode_deltas.size(); index++) { + uncomp_writer.WriteBit(update_loop_filter_mode_deltas[index]); + + if (update_loop_filter_mode_deltas[index]) { + uncomp_writer.WriteS(current_frame_info.mode_deltas[index], 6); + } + } + // save new deltas + loop_filter_ref_deltas = current_frame_info.ref_deltas; + loop_filter_mode_deltas = current_frame_info.mode_deltas; + } + } + + uncomp_writer.WriteU(current_frame_info.base_q_index, 8); + + uncomp_writer.WriteDeltaQ(current_frame_info.y_dc_delta_q); + uncomp_writer.WriteDeltaQ(current_frame_info.uv_dc_delta_q); + uncomp_writer.WriteDeltaQ(current_frame_info.uv_ac_delta_q); + + uncomp_writer.WriteBit(false); // Segmentation enabled (TODO). + + const s32 min_tile_cols_log2 = CalcMinLog2TileCols(current_frame_info.frame_size.width); + const s32 max_tile_cols_log2 = CalcMaxLog2TileCols(current_frame_info.frame_size.width); + + const s32 tile_cols_log2_diff = current_frame_info.log2_tile_cols - min_tile_cols_log2; + const s32 tile_cols_log2_inc_mask = (1 << tile_cols_log2_diff) - 1; + + // If it's less than the maximum, we need to add an extra 0 on the bitstream + // to indicate that it should stop reading. + if (current_frame_info.log2_tile_cols < max_tile_cols_log2) { + uncomp_writer.WriteU(tile_cols_log2_inc_mask << 1, tile_cols_log2_diff + 1); + } else { + uncomp_writer.WriteU(tile_cols_log2_inc_mask, tile_cols_log2_diff); + } + + const bool tile_rows_log2_is_nonzero = current_frame_info.log2_tile_rows != 0; + + uncomp_writer.WriteBit(tile_rows_log2_is_nonzero); + + if (tile_rows_log2_is_nonzero) { + uncomp_writer.WriteBit(current_frame_info.log2_tile_rows > 1); + } + + return uncomp_writer; +} + +std::vector& VP9::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state) { + std::vector bitstream; + { + Vp9FrameContainer curr_frame = GetCurrentFrame(state); + current_frame_info = curr_frame.info; + bitstream = curr_frame.bit_stream; + } + + // The uncompressed header routine sets PrevProb parameters needed for the compressed header + auto uncomp_writer = ComposeUncompressedHeader(); + std::vector compressed_header = ComposeCompressedHeader(); + + uncomp_writer.WriteU(static_cast(compressed_header.size()), 16); + uncomp_writer.Flush(); + std::vector uncompressed_header = uncomp_writer.GetByteArray(); + + // Write headers and frame to buffer + frame.resize(uncompressed_header.size() + compressed_header.size() + bitstream.size()); + std::memcpy(frame.data(), uncompressed_header.data(), uncompressed_header.size()); + std::memcpy(frame.data() + uncompressed_header.size(), compressed_header.data(), + compressed_header.size()); + std::memcpy(frame.data() + uncompressed_header.size() + compressed_header.size(), + bitstream.data(), bitstream.size()); + + // keep track of frame number + current_frame_number++; + grace_period--; + + // don't display hidden frames + hidden = !current_frame_info.show_frame; + return frame; +} + +VpxRangeEncoder::VpxRangeEncoder() { + Write(false); +} + +VpxRangeEncoder::~VpxRangeEncoder() = default; + +void VpxRangeEncoder::Write(s32 value, s32 value_size) { + for (s32 bit = value_size - 1; bit >= 0; bit--) { + Write(((value >> bit) & 1) != 0); + } +} + +void VpxRangeEncoder::Write(bool bit) { + Write(bit, half_probability); +} + +void VpxRangeEncoder::Write(bool bit, s32 probability) { + u32 local_range = range; + const u32 split = 1 + (((local_range - 1) * static_cast(probability)) >> 8); + local_range = split; + + if (bit) { + low_value += split; + local_range = range - split; + } + + s32 shift = norm_lut[local_range]; + local_range <<= shift; + count += shift; + + if (count >= 0) { + const s32 offset = shift - count; + + if (((low_value << (offset - 1)) >> 31) != 0) { + const s32 current_pos = static_cast(base_stream.GetPosition()); + base_stream.Seek(-1, Common::SeekOrigin::FromCurrentPos); + while (base_stream.GetPosition() >= 0 && PeekByte() == 0xff) { + base_stream.WriteByte(0); + + base_stream.Seek(-2, Common::SeekOrigin::FromCurrentPos); + } + base_stream.WriteByte(static_cast((PeekByte() + 1))); + base_stream.Seek(current_pos, Common::SeekOrigin::SetOrigin); + } + base_stream.WriteByte(static_cast((low_value >> (24 - offset)))); + + low_value <<= offset; + shift = count; + low_value &= 0xffffff; + count -= 8; + } + + low_value <<= shift; + range = local_range; +} + +void VpxRangeEncoder::End() { + for (std::size_t index = 0; index < 32; ++index) { + Write(false); + } +} + +u8 VpxRangeEncoder::PeekByte() { + const u8 value = base_stream.ReadByte(); + base_stream.Seek(-1, Common::SeekOrigin::FromCurrentPos); + + return value; +} + +VpxBitStreamWriter::VpxBitStreamWriter() = default; + +VpxBitStreamWriter::~VpxBitStreamWriter() = default; + +void VpxBitStreamWriter::WriteU(u32 value, u32 value_size) { + WriteBits(value, value_size); +} + +void VpxBitStreamWriter::WriteS(s32 value, u32 value_size) { + const bool sign = value < 0; + if (sign) { + value = -value; + } + + WriteBits(static_cast(value << 1) | (sign ? 1 : 0), value_size + 1); +} + +void VpxBitStreamWriter::WriteDeltaQ(u32 value) { + const bool delta_coded = value != 0; + WriteBit(delta_coded); + + if (delta_coded) { + WriteBits(value, 4); + } +} + +void VpxBitStreamWriter::WriteBits(u32 value, u32 bit_count) { + s32 value_pos = 0; + s32 remaining = bit_count; + + while (remaining > 0) { + s32 copy_size = remaining; + + const s32 free = GetFreeBufferBits(); + + if (copy_size > free) { + copy_size = free; + } + + const s32 mask = (1 << copy_size) - 1; + + const s32 src_shift = (bit_count - value_pos) - copy_size; + const s32 dst_shift = (buffer_size - buffer_pos) - copy_size; + + buffer |= ((value >> src_shift) & mask) << dst_shift; + + value_pos += copy_size; + buffer_pos += copy_size; + remaining -= copy_size; + } +} + +void VpxBitStreamWriter::WriteBit(bool state) { + WriteBits(state ? 1 : 0, 1); +} + +s32 VpxBitStreamWriter::GetFreeBufferBits() { + if (buffer_pos == buffer_size) { + Flush(); + } + + return buffer_size - buffer_pos; +} + +void VpxBitStreamWriter::Flush() { + if (buffer_pos == 0) { + return; + } + byte_array.push_back(static_cast(buffer)); + buffer = 0; + buffer_pos = 0; +} + +std::vector& VpxBitStreamWriter::GetByteArray() { + return byte_array; +} + +const std::vector& VpxBitStreamWriter::GetByteArray() const { + return byte_array; +} + +} // namespace Tegra::Decoder -- cgit v1.2.3 From 31d9c3c75e444ca29a5eb58ea92593a0d62a6545 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Tue, 27 Oct 2020 02:09:17 -0400 Subject: nvdec: Tidy up header includes Prevents a few unnecessary inclusions. --- src/video_core/command_classes/codecs/codec.cpp | 1 + src/video_core/command_classes/codecs/codec.h | 2 -- src/video_core/command_classes/codecs/h264.cpp | 15 +++++++++++ src/video_core/command_classes/codecs/h264.h | 12 --------- src/video_core/command_classes/codecs/vp9.cpp | 31 +++++++++++++++++++++- src/video_core/command_classes/codecs/vp9.h | 32 ++--------------------- src/video_core/command_classes/codecs/vp9_types.h | 6 ++--- src/video_core/command_classes/host1x.cpp | 2 +- src/video_core/command_classes/host1x.h | 2 +- src/video_core/command_classes/nvdec.cpp | 6 +---- src/video_core/command_classes/nvdec.h | 6 ++--- src/video_core/command_classes/vic.cpp | 2 +- src/video_core/command_classes/vic.h | 4 +-- 13 files changed, 59 insertions(+), 62 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/codec.cpp b/src/video_core/command_classes/codecs/codec.cpp index 2df410be8f..1adf3cd135 100644 --- a/src/video_core/command_classes/codecs/codec.cpp +++ b/src/video_core/command_classes/codecs/codec.cpp @@ -4,6 +4,7 @@ #include #include +#include #include "common/assert.h" #include "video_core/command_classes/codecs/codec.h" #include "video_core/command_classes/codecs/h264.h" diff --git a/src/video_core/command_classes/codecs/codec.h b/src/video_core/command_classes/codecs/codec.h index 2e56daf29b..cb67094f68 100644 --- a/src/video_core/command_classes/codecs/codec.h +++ b/src/video_core/command_classes/codecs/codec.h @@ -5,8 +5,6 @@ #pragma once #include -#include -#include "common/common_funcs.h" #include "common/common_types.h" #include "video_core/command_classes/nvdec_common.h" diff --git a/src/video_core/command_classes/codecs/h264.cpp b/src/video_core/command_classes/codecs/h264.cpp index 1a39f7b232..54a749e2bd 100644 --- a/src/video_core/command_classes/codecs/h264.cpp +++ b/src/video_core/command_classes/codecs/h264.cpp @@ -18,12 +18,27 @@ // OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. // +#include #include "common/bit_util.h" #include "video_core/command_classes/codecs/h264.h" #include "video_core/gpu.h" #include "video_core/memory_manager.h" namespace Tegra::Decoder { +namespace { +// ZigZag LUTs from libavcodec. +constexpr std::array zig_zag_direct{ + 0, 1, 8, 16, 9, 2, 3, 10, 17, 24, 32, 25, 18, 11, 4, 5, 12, 19, 26, 33, 40, 48, + 41, 34, 27, 20, 13, 6, 7, 14, 21, 28, 35, 42, 49, 56, 57, 50, 43, 36, 29, 22, 15, 23, + 30, 37, 44, 51, 58, 59, 52, 45, 38, 31, 39, 46, 53, 60, 61, 54, 47, 55, 62, 63, +}; + +constexpr std::array zig_zag_scan{ + 0 + 0 * 4, 1 + 0 * 4, 0 + 1 * 4, 0 + 2 * 4, 1 + 1 * 4, 2 + 0 * 4, 3 + 0 * 4, 2 + 1 * 4, + 1 + 2 * 4, 0 + 3 * 4, 1 + 3 * 4, 2 + 2 * 4, 3 + 1 * 4, 3 + 2 * 4, 2 + 3 * 4, 3 + 3 * 4, +}; +} // Anonymous namespace + H264::H264(GPU& gpu_) : gpu(gpu_) {} H264::~H264() = default; diff --git a/src/video_core/command_classes/codecs/h264.h b/src/video_core/command_classes/codecs/h264.h index 21752dd90c..c36a54399b 100644 --- a/src/video_core/command_classes/codecs/h264.h +++ b/src/video_core/command_classes/codecs/h264.h @@ -55,18 +55,6 @@ public: const std::vector& GetByteArray() const; private: - // ZigZag LUTs from libavcodec. - static constexpr std::array zig_zag_direct{ - 0, 1, 8, 16, 9, 2, 3, 10, 17, 24, 32, 25, 18, 11, 4, 5, 12, 19, 26, 33, 40, 48, - 41, 34, 27, 20, 13, 6, 7, 14, 21, 28, 35, 42, 49, 56, 57, 50, 43, 36, 29, 22, 15, 23, - 30, 37, 44, 51, 58, 59, 52, 45, 38, 31, 39, 46, 53, 60, 61, 54, 47, 55, 62, 63, - }; - - static constexpr std::array zig_zag_scan{ - 0 + 0 * 4, 1 + 0 * 4, 0 + 1 * 4, 0 + 2 * 4, 1 + 1 * 4, 2 + 0 * 4, 3 + 0 * 4, 2 + 1 * 4, - 1 + 2 * 4, 0 + 3 * 4, 1 + 3 * 4, 2 + 2 * 4, 3 + 1 * 4, 3 + 2 * 4, 2 + 3 * 4, 3 + 3 * 4, - }; - void WriteBits(s32 value, s32 bit_count); void WriteExpGolombCodedInt(s32 value); void WriteExpGolombCodedUInt(u32 value); diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 3bae0bb5d9..d888e773aa 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -9,7 +9,7 @@ #include "video_core/memory_manager.h" namespace Tegra::Decoder { - +namespace { // Default compressed header probabilities once frame context resets constexpr Vp9EntropyProbs default_probs{ .y_mode_prob{ @@ -170,6 +170,35 @@ constexpr Vp9EntropyProbs default_probs{ .high_precision{128, 128}, }; +constexpr std::array norm_lut{ + 0, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, +}; + +constexpr std::array map_lut{ + 20, 21, 22, 23, 24, 25, 0, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, + 1, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 2, 50, 51, 52, 53, 54, + 55, 56, 57, 58, 59, 60, 61, 3, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, + 73, 4, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 5, 86, 87, 88, 89, + 90, 91, 92, 93, 94, 95, 96, 97, 6, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, + 108, 109, 7, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 8, 122, 123, 124, + 125, 126, 127, 128, 129, 130, 131, 132, 133, 9, 134, 135, 136, 137, 138, 139, 140, 141, 142, + 143, 144, 145, 10, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 11, 158, 159, + 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 12, 170, 171, 172, 173, 174, 175, 176, 177, + 178, 179, 180, 181, 13, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 14, 194, + 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 15, 206, 207, 208, 209, 210, 211, 212, + 213, 214, 215, 216, 217, 16, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 17, + 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 18, 242, 243, 244, 245, 246, 247, + 248, 249, 250, 251, 252, 253, 19, +}; +} // Anonymous namespace + VP9::VP9(GPU& gpu) : gpu(gpu) {} VP9::~VP9() = default; diff --git a/src/video_core/command_classes/codecs/vp9.h b/src/video_core/command_classes/codecs/vp9.h index 748e11baef..dc52ddbde3 100644 --- a/src/video_core/command_classes/codecs/vp9.h +++ b/src/video_core/command_classes/codecs/vp9.h @@ -4,9 +4,9 @@ #pragma once -#include +#include #include -#include "common/common_funcs.h" + #include "common/common_types.h" #include "common/stream.h" #include "video_core/command_classes/codecs/vp9_types.h" @@ -52,17 +52,6 @@ private: u32 range{0xff}; s32 count{-24}; s32 half_probability{128}; - static constexpr std::array norm_lut{ - 0, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, - 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - }; }; class VpxBitStreamWriter { @@ -193,23 +182,6 @@ private: s32 diff_update_probability = 252; s32 frame_sync_code = 0x498342; - static constexpr std::array map_lut = { - 20, 21, 22, 23, 24, 25, 0, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, - 36, 37, 1, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 2, 50, - 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 3, 62, 63, 64, 65, 66, - 67, 68, 69, 70, 71, 72, 73, 4, 74, 75, 76, 77, 78, 79, 80, 81, 82, - 83, 84, 85, 5, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 6, - 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 7, 110, 111, 112, 113, - 114, 115, 116, 117, 118, 119, 120, 121, 8, 122, 123, 124, 125, 126, 127, 128, 129, - 130, 131, 132, 133, 9, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, - 10, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 11, 158, 159, 160, - 161, 162, 163, 164, 165, 166, 167, 168, 169, 12, 170, 171, 172, 173, 174, 175, 176, - 177, 178, 179, 180, 181, 13, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, - 193, 14, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 15, 206, 207, - 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 16, 218, 219, 220, 221, 222, 223, - 224, 225, 226, 227, 228, 229, 17, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, - 240, 241, 18, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 19, - }; }; } // namespace Decoder diff --git a/src/video_core/command_classes/codecs/vp9_types.h b/src/video_core/command_classes/codecs/vp9_types.h index 8688fdac01..a50acf6e89 100644 --- a/src/video_core/command_classes/codecs/vp9_types.h +++ b/src/video_core/command_classes/codecs/vp9_types.h @@ -4,13 +4,11 @@ #pragma once -#include -#include +#include +#include #include -#include "common/cityhash.h" #include "common/common_funcs.h" #include "common/common_types.h" -#include "video_core/command_classes/nvdec_common.h" namespace Tegra { class GPU; diff --git a/src/video_core/command_classes/host1x.cpp b/src/video_core/command_classes/host1x.cpp index a5234ee477..c4dd4881ae 100644 --- a/src/video_core/command_classes/host1x.cpp +++ b/src/video_core/command_classes/host1x.cpp @@ -15,7 +15,7 @@ void Tegra::Host1x::StateWrite(u32 offset, u32 arguments) { std::memcpy(state_offset, &arguments, sizeof(u32)); } -void Tegra::Host1x::ProcessMethod(Host1x::Method method, const std::vector& arguments) { +void Tegra::Host1x::ProcessMethod(Method method, const std::vector& arguments) { StateWrite(static_cast(method), arguments[0]); switch (method) { case Method::WaitSyncpt: diff --git a/src/video_core/command_classes/host1x.h b/src/video_core/command_classes/host1x.h index 501a5ed2ec..013eaa0c12 100644 --- a/src/video_core/command_classes/host1x.h +++ b/src/video_core/command_classes/host1x.h @@ -61,7 +61,7 @@ public: ~Host1x(); /// Writes the method into the state, Invoke Execute() if encountered - void ProcessMethod(Host1x::Method method, const std::vector& arguments); + void ProcessMethod(Method method, const std::vector& arguments); private: /// For Host1x, execute is waiting on a syncpoint previously written into the state diff --git a/src/video_core/command_classes/nvdec.cpp b/src/video_core/command_classes/nvdec.cpp index ede9466ebe..8ca7a7b06f 100644 --- a/src/video_core/command_classes/nvdec.cpp +++ b/src/video_core/command_classes/nvdec.cpp @@ -2,13 +2,9 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include #include "common/assert.h" -#include "common/bit_util.h" -#include "core/memory.h" #include "video_core/command_classes/nvdec.h" #include "video_core/gpu.h" -#include "video_core/memory_manager.h" namespace Tegra { @@ -16,7 +12,7 @@ Nvdec::Nvdec(GPU& gpu_) : gpu(gpu_), codec(std::make_unique(gpu)) {} Nvdec::~Nvdec() = default; -void Nvdec::ProcessMethod(Nvdec::Method method, const std::vector& arguments) { +void Nvdec::ProcessMethod(Method method, const std::vector& arguments) { if (method == Method::SetVideoCodec) { codec->StateWrite(static_cast(method), arguments[0]); } else { diff --git a/src/video_core/command_classes/nvdec.h b/src/video_core/command_classes/nvdec.h index c1a9d843e6..af14f9857b 100644 --- a/src/video_core/command_classes/nvdec.h +++ b/src/video_core/command_classes/nvdec.h @@ -4,8 +4,8 @@ #pragma once +#include #include -#include "common/common_funcs.h" #include "common/common_types.h" #include "video_core/command_classes/codecs/codec.h" @@ -23,7 +23,7 @@ public: ~Nvdec(); /// Writes the method into the state, Invoke Execute() if encountered - void ProcessMethod(Nvdec::Method method, const std::vector& arguments); + void ProcessMethod(Method method, const std::vector& arguments); /// Return most recently decoded frame AVFrame* GetFrame(); @@ -34,6 +34,6 @@ private: void Execute(); GPU& gpu; - std::unique_ptr codec; + std::unique_ptr codec; }; } // namespace Tegra diff --git a/src/video_core/command_classes/vic.cpp b/src/video_core/command_classes/vic.cpp index 66e15a1a82..5b52da277f 100644 --- a/src/video_core/command_classes/vic.cpp +++ b/src/video_core/command_classes/vic.cpp @@ -26,7 +26,7 @@ void Vic::VicStateWrite(u32 offset, u32 arguments) { std::memcpy(state_offset, &arguments, sizeof(u32)); } -void Vic::ProcessMethod(Vic::Method method, const std::vector& arguments) { +void Vic::ProcessMethod(Method method, const std::vector& arguments) { LOG_DEBUG(HW_GPU, "Vic method 0x{:X}", static_cast(method)); VicStateWrite(static_cast(method), arguments[0]); const u64 arg = static_cast(arguments[0]) << 8; diff --git a/src/video_core/command_classes/vic.h b/src/video_core/command_classes/vic.h index dd0a2aed8a..8c4e284a1b 100644 --- a/src/video_core/command_classes/vic.h +++ b/src/video_core/command_classes/vic.h @@ -63,11 +63,11 @@ public: SetOutputSurfaceChromaVOffset = 0x1ca }; - explicit Vic(GPU& gpu, std::shared_ptr nvdec_processor); + explicit Vic(GPU& gpu, std::shared_ptr nvdec_processor); ~Vic(); /// Write to the device state. - void ProcessMethod(Vic::Method method, const std::vector& arguments); + void ProcessMethod(Method method, const std::vector& arguments); private: void Execute(); -- cgit v1.2.3 From 4e10e539d8c1e4c565a6f2f8824de3a1da4a1d96 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Tue, 27 Oct 2020 02:20:17 -0400 Subject: vp9: Resolve variable shadowing --- src/video_core/command_classes/codecs/vp9.cpp | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index d888e773aa..04e015ae8b 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -415,7 +415,7 @@ Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) frame.info.bitstream_size); } // Buffer two frames, saving the last show frame info - if (next_next_frame.bit_stream.size() != 0) { + if (!next_next_frame.bit_stream.empty()) { Vp9FrameContainer temp{ .info = frame.info, .bit_stream = frame.bit_stream, @@ -425,15 +425,15 @@ Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) frame.bit_stream = next_next_frame.bit_stream; next_next_frame = std::move(temp); - if (next_frame.bit_stream.size() != 0) { - Vp9FrameContainer temp{ + if (!next_frame.bit_stream.empty()) { + Vp9FrameContainer temp2{ .info = frame.info, .bit_stream = frame.bit_stream, }; next_frame.info.show_frame = frame.info.last_frame_shown; frame.info = next_frame.info; frame.bit_stream = next_frame.bit_stream; - next_frame = std::move(temp); + next_frame = std::move(temp2); } else { next_frame.info = frame.info; next_frame.bit_stream = frame.bit_stream; -- cgit v1.2.3 From 9efecc310bb9130d2522e1de7d528144bf2559fb Mon Sep 17 00:00:00 2001 From: Lioncash Date: Tue, 27 Oct 2020 02:20:41 -0400 Subject: vp9: Remove pessimizing moves The move will already occur without std::move. --- src/video_core/command_classes/codecs/vp9.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 04e015ae8b..0315a80bce 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -395,7 +395,7 @@ Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) // to avoid buffering frame data needed for reference frame updating in the header composition. std::memcpy(vp9_info.frame_offsets.data(), state.surface_luma_offset.data(), 4 * sizeof(u64)); - return std::move(vp9_info); + return vp9_info; } void VP9::InsertEntropy(u64 offset, Vp9EntropyProbs& dst) { @@ -408,7 +408,7 @@ Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) Vp9FrameContainer frame{}; { gpu.SyncGuestHost(); - frame.info = std::move(GetVp9PictureInfo(state)); + frame.info = GetVp9PictureInfo(state); frame.bit_stream.resize(frame.info.bitstream_size); gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, frame.bit_stream.data(), -- cgit v1.2.3 From 617f0167467607be7b6931182e744073b48de1ea Mon Sep 17 00:00:00 2001 From: Lioncash Date: Tue, 27 Oct 2020 02:26:03 -0400 Subject: vp9: Join declarations with assignments --- src/video_core/command_classes/codecs/vp9.cpp | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 0315a80bce..3801809752 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -677,7 +677,6 @@ VpxBitStreamWriter VP9::ComposeUncompressedHeader() { current_frame_info.intra_only = true; } else { - std::array ref_frame_index; if (!current_frame_info.show_frame) { uncomp_writer.WriteBit(current_frame_info.intra_only); @@ -692,9 +691,9 @@ VpxBitStreamWriter VP9::ComposeUncompressedHeader() { } // Last, Golden, Altref frames - ref_frame_index = std::array{0, 1, 2}; + std::array ref_frame_index{0, 1, 2}; - // set when next frame is hidden + // Set when next frame is hidden // altref and golden references are swapped if (swap_next_golden) { ref_frame_index = std::array{0, 2, 1}; @@ -783,17 +782,19 @@ VpxBitStreamWriter VP9::ComposeUncompressedHeader() { for (std::size_t index = 0; index < current_frame_info.ref_deltas.size(); index++) { const s8 old_deltas = loop_filter_ref_deltas[index]; const s8 new_deltas = current_frame_info.ref_deltas[index]; + const bool differing_delta = old_deltas != new_deltas; - loop_filter_delta_update |= - (update_loop_filter_ref_deltas[index] = old_deltas != new_deltas); + update_loop_filter_ref_deltas[index] = differing_delta; + loop_filter_delta_update |= differing_delta; } for (std::size_t index = 0; index < current_frame_info.mode_deltas.size(); index++) { const s8 old_deltas = loop_filter_mode_deltas[index]; const s8 new_deltas = current_frame_info.mode_deltas[index]; + const bool differing_delta = old_deltas != new_deltas; - loop_filter_delta_update |= - (update_loop_filter_mode_deltas[index] = old_deltas != new_deltas); + update_loop_filter_mode_deltas[index] = differing_delta; + loop_filter_delta_update |= differing_delta; } uncomp_writer.WriteBit(loop_filter_delta_update); -- cgit v1.2.3 From c0992fd760ac4ba9ea43212c6e8399ac190ab3d9 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Tue, 27 Oct 2020 02:26:17 -0400 Subject: vp9: Remove dead code --- src/video_core/command_classes/codecs/vp9.cpp | 6 ------ 1 file changed, 6 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 3801809752..2a3bdb7e7d 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -634,12 +634,6 @@ std::vector VP9::ComposeCompressedHeader() { writer.End(); return writer.GetBuffer(); - - const auto writer_bytearray = writer.GetBuffer(); - - std::vector compressed_header(writer_bytearray.size()); - std::memcpy(compressed_header.data(), writer_bytearray.data(), writer_bytearray.size()); - return compressed_header; } VpxBitStreamWriter VP9::ComposeUncompressedHeader() { -- cgit v1.2.3 From 105a345f4e5b003e6710598873674d4bb245afb3 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Tue, 27 Oct 2020 02:27:29 -0400 Subject: vp9: std::move buffer within ComposeFrameHeader() We can move the buffer here to avoid a heap reallocation --- src/video_core/command_classes/codecs/vp9.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 2a3bdb7e7d..bcbea9c054 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -853,7 +853,7 @@ std::vector& VP9::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state) { { Vp9FrameContainer curr_frame = GetCurrentFrame(state); current_frame_info = curr_frame.info; - bitstream = curr_frame.bit_stream; + bitstream = std::move(curr_frame.bit_stream); } // The uncompressed header routine sets PrevProb parameters needed for the compressed header -- cgit v1.2.3 From 1ae83819d9cab957a24b7654f7835269aa1420eb Mon Sep 17 00:00:00 2001 From: ReinUsesLisp Date: Wed, 28 Oct 2020 02:31:30 -0300 Subject: video_core: Enforce -Werror=type-limits Silences one warning and avoids introducing more in the future. --- src/video_core/CMakeLists.txt | 1 + src/video_core/command_classes/codecs/vp9.cpp | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index fdfc885fc4..15bd92f098 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -303,6 +303,7 @@ else() -Werror=conversion -Wno-error=sign-conversion -Werror=switch + -Werror=type-limits -Werror=unused-variable $<$:-Werror=class-memaccess> diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 3bae0bb5d9..c47a0e0601 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -893,7 +893,7 @@ void VpxRangeEncoder::Write(bool bit, s32 probability) { if (((low_value << (offset - 1)) >> 31) != 0) { const s32 current_pos = static_cast(base_stream.GetPosition()); base_stream.Seek(-1, Common::SeekOrigin::FromCurrentPos); - while (base_stream.GetPosition() >= 0 && PeekByte() == 0xff) { + while (PeekByte() == 0xff) { base_stream.WriteByte(0); base_stream.Seek(-2, Common::SeekOrigin::FromCurrentPos); -- cgit v1.2.3 From de16b5a409cb4e9d3cce147d6fdce1fd46407da7 Mon Sep 17 00:00:00 2001 From: ReinUsesLisp Date: Wed, 28 Oct 2020 02:25:40 -0300 Subject: video_core: Enforce -Wredundant-move and -Wpessimizing-move Silence three warnings and make them errors to avoid introducing more in the future. --- src/video_core/CMakeLists.txt | 2 ++ src/video_core/command_classes/codecs/vp9.cpp | 2 +- src/video_core/renderer_opengl/gl_shader_disk_cache.cpp | 3 +-- src/video_core/renderer_vulkan/wrapper.cpp | 2 +- 4 files changed, 5 insertions(+), 4 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index fdfc885fc4..5670842410 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -302,6 +302,8 @@ else() target_compile_options(video_core PRIVATE -Werror=conversion -Wno-error=sign-conversion + -Werror=pessimizing-move + -Werror=redundant-move -Werror=switch -Werror=unused-variable diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 3bae0bb5d9..747c63d85a 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -366,7 +366,7 @@ Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) // to avoid buffering frame data needed for reference frame updating in the header composition. std::memcpy(vp9_info.frame_offsets.data(), state.surface_luma_offset.data(), 4 * sizeof(u64)); - return std::move(vp9_info); + return vp9_info; } void VP9::InsertEntropy(u64 offset, Vp9EntropyProbs& dst) { diff --git a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp index 166ee34e14..70dd0c3c64 100644 --- a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp @@ -317,8 +317,7 @@ std::optional> ShaderDiskCacheOpenGL::Lo return std::nullopt; } } - - return std::move(entries); + return entries; } void ShaderDiskCacheOpenGL::InvalidateTransferable() { diff --git a/src/video_core/renderer_vulkan/wrapper.cpp b/src/video_core/renderer_vulkan/wrapper.cpp index c034558a3f..4e83303d89 100644 --- a/src/video_core/renderer_vulkan/wrapper.cpp +++ b/src/video_core/renderer_vulkan/wrapper.cpp @@ -844,7 +844,7 @@ std::optional> EnumerateInstanceExtensionProp VK_SUCCESS) { return std::nullopt; } - return std::move(properties); + return properties; } std::optional> EnumerateInstanceLayerProperties( -- cgit v1.2.3 From e54c1f4a41816a5b0422a81715095c8dc87ac38a Mon Sep 17 00:00:00 2001 From: Lioncash Date: Thu, 29 Oct 2020 22:34:44 -0400 Subject: vp9: Make some member functions internally linked These helper functions don't directly modify any member state and can be hidden from view. --- src/video_core/command_classes/codecs/vp9.cpp | 102 ++++++++++++++------------ src/video_core/command_classes/codecs/vp9.h | 10 --- 2 files changed, 54 insertions(+), 58 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index b3e98aa9f2..aeb9866ded 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -197,6 +197,60 @@ constexpr std::array map_lut{ 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 18, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 19, }; + +// 6.2.14 Tile size calculation + +s32 CalcMinLog2TileCols(s32 frame_width) { + const s32 sb64_cols = (frame_width + 63) / 64; + s32 min_log2 = 0; + + while ((64 << min_log2) < sb64_cols) { + min_log2++; + } + + return min_log2; +} + +s32 CalcMaxLog2TileCols(s32 frame_width) { + const s32 sb64_cols = (frame_width + 63) / 64; + s32 max_log2 = 1; + + while ((sb64_cols >> max_log2) >= 4) { + max_log2++; + } + + return max_log2 - 1; +} + +// Recenters probability. Based on section 6.3.6 of VP9 Specification +s32 RecenterNonNeg(s32 new_prob, s32 old_prob) { + if (new_prob > old_prob * 2) { + return new_prob; + } + + if (new_prob >= old_prob) { + return (new_prob - old_prob) * 2; + } + + return (old_prob - new_prob) * 2 - 1; +} + +// Adjusts old_prob depending on new_prob. Based on section 6.3.5 of VP9 Specification +s32 RemapProbability(s32 new_prob, s32 old_prob) { + new_prob--; + old_prob--; + + std::size_t index{}; + + if (old_prob * 2 <= 0xff) { + index = static_cast(std::max(0, RecenterNonNeg(new_prob, old_prob) - 1)); + } else { + index = static_cast( + std::max(0, RecenterNonNeg(0xff - 1 - new_prob, 0xff - 1 - old_prob) - 1)); + } + + return map_lut[index]; +} } // Anonymous namespace VP9::VP9(GPU& gpu) : gpu(gpu) {} @@ -236,32 +290,6 @@ void VP9::WriteProbabilityDelta(VpxRangeEncoder& writer, u8 new_prob, u8 old_pro EncodeTermSubExp(writer, delta); } -s32 VP9::RemapProbability(s32 new_prob, s32 old_prob) { - new_prob--; - old_prob--; - - std::size_t index{}; - - if (old_prob * 2 <= 0xff) { - index = static_cast(std::max(0, RecenterNonNeg(new_prob, old_prob) - 1)); - } else { - index = static_cast( - std::max(0, RecenterNonNeg(0xff - 1 - new_prob, 0xff - 1 - old_prob) - 1)); - } - - return map_lut[index]; -} - -s32 VP9::RecenterNonNeg(s32 new_prob, s32 old_prob) { - if (new_prob > old_prob * 2) { - return new_prob; - } else if (new_prob >= old_prob) { - return (new_prob - old_prob) * 2; - } else { - return (old_prob - new_prob) * 2 - 1; - } -} - void VP9::EncodeTermSubExp(VpxRangeEncoder& writer, s32 value) { if (WriteLessThan(writer, value, 16)) { writer.Write(value, 4); @@ -361,28 +389,6 @@ void VP9::WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_ } } -s32 VP9::CalcMinLog2TileCols(s32 frame_width) { - const s32 sb64_cols = (frame_width + 63) / 64; - s32 min_log2 = 0; - - while ((64 << min_log2) < sb64_cols) { - min_log2++; - } - - return min_log2; -} - -s32 VP9::CalcMaxLog2TileCols(s32 frameWidth) { - const s32 sb64_cols = (frameWidth + 63) / 64; - s32 max_log2 = 1; - - while ((sb64_cols >> max_log2) >= 4) { - max_log2++; - } - - return max_log2 - 1; -} - Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) { PictureInfo picture_info{}; gpu.MemoryManager().ReadBlock(state.picture_info_offset, &picture_info, sizeof(PictureInfo)); diff --git a/src/video_core/command_classes/codecs/vp9.h b/src/video_core/command_classes/codecs/vp9.h index dc52ddbde3..3826f2c955 100644 --- a/src/video_core/command_classes/codecs/vp9.h +++ b/src/video_core/command_classes/codecs/vp9.h @@ -121,12 +121,6 @@ private: /// Generates compressed header probability deltas in the bitstream writer void WriteProbabilityDelta(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); - /// Adjusts old_prob depending on new_prob. Based on section 6.3.5 of VP9 Specification - s32 RemapProbability(s32 new_prob, s32 old_prob); - - /// Recenters probability. Based on section 6.3.6 of VP9 Specification - s32 RecenterNonNeg(s32 new_prob, s32 old_prob); - /// Inverse of 6.3.4 Decode term subexp void EncodeTermSubExp(VpxRangeEncoder& writer, s32 value); @@ -146,10 +140,6 @@ private: /// Write motion vector probability updates. 6.3.17 in the spec void WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); - /// 6.2.14 Tile size calculation - s32 CalcMinLog2TileCols(s32 frame_width); - s32 CalcMaxLog2TileCols(s32 frame_width); - /// Returns VP9 information from NVDEC provided offset and size Vp9PictureInfo GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state); -- cgit v1.2.3 From 0c2f517cd06457e0b2378208ed84dfca2e198941 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Thu, 29 Oct 2020 22:40:46 -0400 Subject: vp9: Mark functions with [[nodiscard]] where applicable Prevents values from mistakenly being discarded in cases where it's a bug to do so. --- src/video_core/command_classes/codecs/vp9.cpp | 8 ++++---- src/video_core/command_classes/codecs/vp9.h | 18 +++++++++--------- 2 files changed, 13 insertions(+), 13 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index aeb9866ded..42520f856f 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -200,7 +200,7 @@ constexpr std::array map_lut{ // 6.2.14 Tile size calculation -s32 CalcMinLog2TileCols(s32 frame_width) { +[[nodiscard]] s32 CalcMinLog2TileCols(s32 frame_width) { const s32 sb64_cols = (frame_width + 63) / 64; s32 min_log2 = 0; @@ -211,7 +211,7 @@ s32 CalcMinLog2TileCols(s32 frame_width) { return min_log2; } -s32 CalcMaxLog2TileCols(s32 frame_width) { +[[nodiscard]] s32 CalcMaxLog2TileCols(s32 frame_width) { const s32 sb64_cols = (frame_width + 63) / 64; s32 max_log2 = 1; @@ -223,7 +223,7 @@ s32 CalcMaxLog2TileCols(s32 frame_width) { } // Recenters probability. Based on section 6.3.6 of VP9 Specification -s32 RecenterNonNeg(s32 new_prob, s32 old_prob) { +[[nodiscard]] s32 RecenterNonNeg(s32 new_prob, s32 old_prob) { if (new_prob > old_prob * 2) { return new_prob; } @@ -236,7 +236,7 @@ s32 RecenterNonNeg(s32 new_prob, s32 old_prob) { } // Adjusts old_prob depending on new_prob. Based on section 6.3.5 of VP9 Specification -s32 RemapProbability(s32 new_prob, s32 old_prob) { +[[nodiscard]] s32 RemapProbability(s32 new_prob, s32 old_prob) { new_prob--; old_prob--; diff --git a/src/video_core/command_classes/codecs/vp9.h b/src/video_core/command_classes/codecs/vp9.h index 94e8f90902..76b5a8283c 100644 --- a/src/video_core/command_classes/codecs/vp9.h +++ b/src/video_core/command_classes/codecs/vp9.h @@ -37,11 +37,11 @@ public: /// Signal the end of the bitstream void End(); - std::vector& GetBuffer() { + [[nodiscard]] std::vector& GetBuffer() { return base_stream.GetBuffer(); } - const std::vector& GetBuffer() const { + [[nodiscard]] const std::vector& GetBuffer() const { return base_stream.GetBuffer(); } @@ -75,10 +75,10 @@ public: void Flush(); /// Returns byte_array - std::vector& GetByteArray(); + [[nodiscard]] std::vector& GetByteArray(); /// Returns const byte_array - const std::vector& GetByteArray() const; + [[nodiscard]] const std::vector& GetByteArray() const; private: /// Write bit_count bits from value into buffer @@ -104,7 +104,7 @@ public: std::vector& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state); /// Returns true if the most recent frame was a hidden frame. - bool WasFrameHidden() const { + [[nodiscard]] bool WasFrameHidden() const { return hidden; } @@ -141,17 +141,17 @@ private: void WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); /// Returns VP9 information from NVDEC provided offset and size - Vp9PictureInfo GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state); + [[nodiscard]] Vp9PictureInfo GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state); /// Read and convert NVDEC provided entropy probs to Vp9EntropyProbs struct void InsertEntropy(u64 offset, Vp9EntropyProbs& dst); /// Returns frame to be decoded after buffering - Vp9FrameContainer GetCurrentFrame(const NvdecCommon::NvdecRegisters& state); + [[nodiscard]] Vp9FrameContainer GetCurrentFrame(const NvdecCommon::NvdecRegisters& state); /// Use NVDEC providied information to compose the headers for the current frame - std::vector ComposeCompressedHeader(); - VpxBitStreamWriter ComposeUncompressedHeader(); + [[nodiscard]] std::vector ComposeCompressedHeader(); + [[nodiscard]] VpxBitStreamWriter ComposeUncompressedHeader(); GPU& gpu; std::vector frame; -- cgit v1.2.3 From 2ee59021c29bf6f3fb3b665923bb2b0589ceacf8 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Fri, 30 Oct 2020 16:13:29 -0400 Subject: nvdec: Make use of [[nodiscard]] where applicable Prevents bugs from occurring where the results of a function are accidentally discarded --- src/video_core/command_classes/codecs/codec.h | 6 +++--- src/video_core/command_classes/codecs/h264.cpp | 3 ++- src/video_core/command_classes/codecs/h264.h | 10 +++++----- src/video_core/command_classes/codecs/vp9.cpp | 2 +- src/video_core/command_classes/codecs/vp9.h | 2 +- src/video_core/command_classes/codecs/vp9_types.h | 5 ++--- src/video_core/command_classes/nvdec.h | 4 ++-- 7 files changed, 16 insertions(+), 16 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/codec.h b/src/video_core/command_classes/codecs/codec.h index cb67094f68..5bbe6a332f 100644 --- a/src/video_core/command_classes/codecs/codec.h +++ b/src/video_core/command_classes/codecs/codec.h @@ -42,11 +42,11 @@ public: void Decode(); /// Returns most recently decoded frame - AVFrame* GetCurrentFrame(); - const AVFrame* GetCurrentFrame() const; + [[nodiscard]] AVFrame* GetCurrentFrame(); + [[nodiscard]] const AVFrame* GetCurrentFrame() const; /// Returns the value of current_codec - NvdecCommon::VideoCodec GetCurrentCodec() const; + [[nodiscard]] NvdecCommon::VideoCodec GetCurrentCodec() const; private: bool initialized{}; diff --git a/src/video_core/command_classes/codecs/h264.cpp b/src/video_core/command_classes/codecs/h264.cpp index 549a40f527..33e063e201 100644 --- a/src/video_core/command_classes/codecs/h264.cpp +++ b/src/video_core/command_classes/codecs/h264.cpp @@ -43,7 +43,8 @@ H264::H264(GPU& gpu_) : gpu(gpu_) {} H264::~H264() = default; -std::vector& H264::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state, bool is_first_frame) { +const std::vector& H264::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state, + bool is_first_frame) { H264DecoderContext context{}; gpu.MemoryManager().ReadBlock(state.picture_info_offset, &context, sizeof(H264DecoderContext)); diff --git a/src/video_core/command_classes/codecs/h264.h b/src/video_core/command_classes/codecs/h264.h index f2292fd2fd..2734494954 100644 --- a/src/video_core/command_classes/codecs/h264.h +++ b/src/video_core/command_classes/codecs/h264.h @@ -51,14 +51,14 @@ public: void WriteScalingList(const std::vector& list, s32 start, s32 count); /// Return the bitstream as a vector. - std::vector& GetByteArray(); - const std::vector& GetByteArray() const; + [[nodiscard]] std::vector& GetByteArray(); + [[nodiscard]] const std::vector& GetByteArray() const; private: void WriteBits(s32 value, s32 bit_count); void WriteExpGolombCodedInt(s32 value); void WriteExpGolombCodedUInt(u32 value); - s32 GetFreeBufferBits(); + [[nodiscard]] s32 GetFreeBufferBits(); void Flush(); s32 buffer_size{8}; @@ -74,8 +74,8 @@ public: ~H264(); /// Compose the H264 header of the frame for FFmpeg decoding - std::vector& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state, - bool is_first_frame = false); + [[nodiscard]] const std::vector& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state, + bool is_first_frame = false); private: struct H264ParameterSet { diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 42520f856f..ab44fdc9e4 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -854,7 +854,7 @@ VpxBitStreamWriter VP9::ComposeUncompressedHeader() { return uncomp_writer; } -std::vector& VP9::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state) { +const std::vector& VP9::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state) { std::vector bitstream; { Vp9FrameContainer curr_frame = GetCurrentFrame(state); diff --git a/src/video_core/command_classes/codecs/vp9.h b/src/video_core/command_classes/codecs/vp9.h index 05c9682fa5..e2504512c7 100644 --- a/src/video_core/command_classes/codecs/vp9.h +++ b/src/video_core/command_classes/codecs/vp9.h @@ -119,7 +119,7 @@ public: /// Composes the VP9 frame from the GPU state information. Based on the official VP9 spec /// documentation - std::vector& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state); + [[nodiscard]] const std::vector& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state); /// Returns true if the most recent frame was a hidden frame. [[nodiscard]] bool WasFrameHidden() const { diff --git a/src/video_core/command_classes/codecs/vp9_types.h b/src/video_core/command_classes/codecs/vp9_types.h index a50acf6e89..4f0b05d225 100644 --- a/src/video_core/command_classes/codecs/vp9_types.h +++ b/src/video_core/command_classes/codecs/vp9_types.h @@ -231,9 +231,8 @@ struct PictureInfo { u32 surface_params{}; INSERT_PADDING_WORDS(3); - Vp9PictureInfo Convert() const { - - return Vp9PictureInfo{ + [[nodiscard]] Vp9PictureInfo Convert() const { + return { .is_key_frame = (vp9_flags & FrameFlags::IsKeyFrame) != 0, .intra_only = (vp9_flags & FrameFlags::IntraOnly) != 0, .last_frame_was_key = (vp9_flags & FrameFlags::LastFrameIsKeyFrame) != 0, diff --git a/src/video_core/command_classes/nvdec.h b/src/video_core/command_classes/nvdec.h index af14f9857b..eec4443f9f 100644 --- a/src/video_core/command_classes/nvdec.h +++ b/src/video_core/command_classes/nvdec.h @@ -26,8 +26,8 @@ public: void ProcessMethod(Method method, const std::vector& arguments); /// Return most recently decoded frame - AVFrame* GetFrame(); - const AVFrame* GetFrame() const; + [[nodiscard]] AVFrame* GetFrame(); + [[nodiscard]] const AVFrame* GetFrame() const; private: /// Invoke codec to decode a frame -- cgit v1.2.3 From 5beb9903491eab98e1a3112a384cbd0d66089b39 Mon Sep 17 00:00:00 2001 From: ameerj Date: Wed, 25 Nov 2020 17:10:44 -0500 Subject: Queue decoded frames, cleanup decoders --- src/video_core/command_classes/codecs/codec.cpp | 30 +- src/video_core/command_classes/codecs/codec.h | 11 +- src/video_core/command_classes/codecs/h264.cpp | 2 +- src/video_core/command_classes/codecs/h264.h | 4 +- src/video_core/command_classes/codecs/vp9.cpp | 333 +++++++++------------- src/video_core/command_classes/codecs/vp9.h | 7 +- src/video_core/command_classes/codecs/vp9_types.h | 154 +++------- src/video_core/command_classes/nvdec.cpp | 6 +- src/video_core/command_classes/nvdec.h | 3 +- src/video_core/command_classes/vic.cpp | 15 +- 10 files changed, 227 insertions(+), 338 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/codec.cpp b/src/video_core/command_classes/codecs/codec.cpp index 1adf3cd135..1a19341c85 100644 --- a/src/video_core/command_classes/codecs/codec.cpp +++ b/src/video_core/command_classes/codecs/codec.cpp @@ -18,6 +18,11 @@ extern "C" { namespace Tegra { +void av_frame_deleter(AVFrame* ptr) { + av_frame_unref(ptr); + av_free(ptr); +} + Codec::Codec(GPU& gpu_) : gpu(gpu_), h264_decoder(std::make_unique(gpu)), vp9_decoder(std::make_unique(gpu)) {} @@ -27,7 +32,9 @@ Codec::~Codec() { return; } // Free libav memory + AVFrame* av_frame{nullptr}; avcodec_send_packet(av_codec_ctx, nullptr); + av_frame = av_frame_alloc(); avcodec_receive_frame(av_codec_ctx, av_frame); avcodec_flush_buffers(av_codec_ctx); @@ -60,7 +67,7 @@ void Codec::Decode() { } av_codec_ctx = avcodec_alloc_context3(av_codec); - av_frame = av_frame_alloc(); + av_codec_ctx->refcounted_frames = 1; av_opt_set(av_codec_ctx->priv_data, "tune", "zerolatency", 0); // TODO(ameerj): libavcodec gpu hw acceleration @@ -68,8 +75,6 @@ void Codec::Decode() { const auto av_error = avcodec_open2(av_codec_ctx, av_codec, nullptr); if (av_error < 0) { LOG_ERROR(Service_NVDRV, "avcodec_open2() Failed."); - av_frame_unref(av_frame); - av_free(av_frame); avcodec_close(av_codec_ctx); return; } @@ -96,16 +101,21 @@ void Codec::Decode() { if (!vp9_hidden_frame) { // Only receive/store visible frames - avcodec_receive_frame(av_codec_ctx, av_frame); + AVFramePtr frame = AVFramePtr{av_frame_alloc(), av_frame_deleter}; + avcodec_receive_frame(av_codec_ctx, frame.get()); + av_frames.push(std::move(frame)); } } -AVFrame* Codec::GetCurrentFrame() { - return av_frame; -} - -const AVFrame* Codec::GetCurrentFrame() const { - return av_frame; +AVFramePtr Codec::GetCurrentFrame() { + // Sometimes VIC will request more frames than have been decoded. + // in this case, return a nullptr and don't overwrite previous frame data + if (av_frames.size() > 0) { + AVFramePtr frame = std::move(av_frames.front()); + av_frames.pop(); + return frame; + } + return AVFramePtr{nullptr, av_frame_deleter}; } NvdecCommon::VideoCodec Codec::GetCurrentCodec() const { diff --git a/src/video_core/command_classes/codecs/codec.h b/src/video_core/command_classes/codecs/codec.h index 5bbe6a332f..c26b59fdec 100644 --- a/src/video_core/command_classes/codecs/codec.h +++ b/src/video_core/command_classes/codecs/codec.h @@ -5,6 +5,7 @@ #pragma once #include +#include #include "common/common_types.h" #include "video_core/command_classes/nvdec_common.h" @@ -22,6 +23,9 @@ namespace Tegra { class GPU; struct VicRegisters; +void av_frame_deleter(AVFrame* ptr); +using AVFramePtr = std::unique_ptr; + namespace Decoder { class H264; class VP9; @@ -41,9 +45,8 @@ public: /// Call decoders to construct headers, decode AVFrame with ffmpeg void Decode(); - /// Returns most recently decoded frame - [[nodiscard]] AVFrame* GetCurrentFrame(); - [[nodiscard]] const AVFrame* GetCurrentFrame() const; + /// Returns next decoded frame + [[nodiscard]] AVFramePtr GetCurrentFrame(); /// Returns the value of current_codec [[nodiscard]] NvdecCommon::VideoCodec GetCurrentCodec() const; @@ -54,13 +57,13 @@ private: AVCodec* av_codec{nullptr}; AVCodecContext* av_codec_ctx{nullptr}; - AVFrame* av_frame{nullptr}; GPU& gpu; std::unique_ptr h264_decoder; std::unique_ptr vp9_decoder; NvdecCommon::NvdecRegisters state{}; + std::queue av_frames{}; }; } // namespace Tegra diff --git a/src/video_core/command_classes/codecs/h264.cpp b/src/video_core/command_classes/codecs/h264.cpp index 33e063e201..65bbeac788 100644 --- a/src/video_core/command_classes/codecs/h264.cpp +++ b/src/video_core/command_classes/codecs/h264.cpp @@ -43,7 +43,7 @@ H264::H264(GPU& gpu_) : gpu(gpu_) {} H264::~H264() = default; -const std::vector& H264::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state, +const std::vector& H264::ComposeFrameHeader(const NvdecCommon::NvdecRegisters& state, bool is_first_frame) { H264DecoderContext context{}; gpu.MemoryManager().ReadBlock(state.picture_info_offset, &context, sizeof(H264DecoderContext)); diff --git a/src/video_core/command_classes/codecs/h264.h b/src/video_core/command_classes/codecs/h264.h index 2734494954..0f3a1d9f3f 100644 --- a/src/video_core/command_classes/codecs/h264.h +++ b/src/video_core/command_classes/codecs/h264.h @@ -74,8 +74,8 @@ public: ~H264(); /// Compose the H264 header of the frame for FFmpeg decoding - [[nodiscard]] const std::vector& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state, - bool is_first_frame = false); + [[nodiscard]] const std::vector& ComposeFrameHeader( + const NvdecCommon::NvdecRegisters& state, bool is_first_frame = false); private: struct H264ParameterSet { diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index ab44fdc9e4..31e00c27d7 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -23,122 +23,102 @@ constexpr Vp9EntropyProbs default_probs{ 222, 34, 30, 0, 72, 16, 44, 0, 58, 32, 12, 0, 10, 7, 6, 0, }, .coef_probs{ - 195, 29, 183, 0, 84, 49, 136, 0, 8, 42, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 31, 107, 169, 0, 35, 99, 159, 0, 17, 82, 140, 0, 8, 66, 114, 0, - 2, 44, 76, 0, 1, 19, 32, 0, 40, 132, 201, 0, 29, 114, 187, 0, 13, 91, 157, 0, - 7, 75, 127, 0, 3, 58, 95, 0, 1, 28, 47, 0, 69, 142, 221, 0, 42, 122, 201, 0, - 15, 91, 159, 0, 6, 67, 121, 0, 1, 42, 77, 0, 1, 17, 31, 0, 102, 148, 228, 0, - 67, 117, 204, 0, 17, 82, 154, 0, 6, 59, 114, 0, 2, 39, 75, 0, 1, 15, 29, 0, - 156, 57, 233, 0, 119, 57, 212, 0, 58, 48, 163, 0, 29, 40, 124, 0, 12, 30, 81, 0, - 3, 12, 31, 0, 191, 107, 226, 0, 124, 117, 204, 0, 25, 99, 155, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 29, 148, 210, 0, 37, 126, 194, 0, 8, 93, 157, 0, - 2, 68, 118, 0, 1, 39, 69, 0, 1, 17, 33, 0, 41, 151, 213, 0, 27, 123, 193, 0, - 3, 82, 144, 0, 1, 58, 105, 0, 1, 32, 60, 0, 1, 13, 26, 0, 59, 159, 220, 0, - 23, 126, 198, 0, 4, 88, 151, 0, 1, 66, 114, 0, 1, 38, 71, 0, 1, 18, 34, 0, - 114, 136, 232, 0, 51, 114, 207, 0, 11, 83, 155, 0, 3, 56, 105, 0, 1, 33, 65, 0, - 1, 17, 34, 0, 149, 65, 234, 0, 121, 57, 215, 0, 61, 49, 166, 0, 28, 36, 114, 0, - 12, 25, 76, 0, 3, 16, 42, 0, 214, 49, 220, 0, 132, 63, 188, 0, 42, 65, 137, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 137, 221, 0, 104, 131, 216, 0, - 49, 111, 192, 0, 21, 87, 155, 0, 2, 49, 87, 0, 1, 16, 28, 0, 89, 163, 230, 0, - 90, 137, 220, 0, 29, 100, 183, 0, 10, 70, 135, 0, 2, 42, 81, 0, 1, 17, 33, 0, - 108, 167, 237, 0, 55, 133, 222, 0, 15, 97, 179, 0, 4, 72, 135, 0, 1, 45, 85, 0, - 1, 19, 38, 0, 124, 146, 240, 0, 66, 124, 224, 0, 17, 88, 175, 0, 4, 58, 122, 0, - 1, 36, 75, 0, 1, 18, 37, 0, 141, 79, 241, 0, 126, 70, 227, 0, 66, 58, 182, 0, - 30, 44, 136, 0, 12, 34, 96, 0, 2, 20, 47, 0, 229, 99, 249, 0, 143, 111, 235, 0, - 46, 109, 192, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 82, 158, 236, 0, - 94, 146, 224, 0, 25, 117, 191, 0, 9, 87, 149, 0, 3, 56, 99, 0, 1, 33, 57, 0, - 83, 167, 237, 0, 68, 145, 222, 0, 10, 103, 177, 0, 2, 72, 131, 0, 1, 41, 79, 0, - 1, 20, 39, 0, 99, 167, 239, 0, 47, 141, 224, 0, 10, 104, 178, 0, 2, 73, 133, 0, - 1, 44, 85, 0, 1, 22, 47, 0, 127, 145, 243, 0, 71, 129, 228, 0, 17, 93, 177, 0, - 3, 61, 124, 0, 1, 41, 84, 0, 1, 21, 52, 0, 157, 78, 244, 0, 140, 72, 231, 0, - 69, 58, 184, 0, 31, 44, 137, 0, 14, 38, 105, 0, 8, 23, 61, 0, 125, 34, 187, 0, - 52, 41, 133, 0, 6, 31, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 37, 109, 153, 0, 51, 102, 147, 0, 23, 87, 128, 0, 8, 67, 101, 0, 1, 41, 63, 0, - 1, 19, 29, 0, 31, 154, 185, 0, 17, 127, 175, 0, 6, 96, 145, 0, 2, 73, 114, 0, - 1, 51, 82, 0, 1, 28, 45, 0, 23, 163, 200, 0, 10, 131, 185, 0, 2, 93, 148, 0, - 1, 67, 111, 0, 1, 41, 69, 0, 1, 14, 24, 0, 29, 176, 217, 0, 12, 145, 201, 0, - 3, 101, 156, 0, 1, 69, 111, 0, 1, 39, 63, 0, 1, 14, 23, 0, 57, 192, 233, 0, - 25, 154, 215, 0, 6, 109, 167, 0, 3, 78, 118, 0, 1, 48, 69, 0, 1, 21, 29, 0, - 202, 105, 245, 0, 108, 106, 216, 0, 18, 90, 144, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 33, 172, 219, 0, 64, 149, 206, 0, 14, 117, 177, 0, 5, 90, 141, 0, - 2, 61, 95, 0, 1, 37, 57, 0, 33, 179, 220, 0, 11, 140, 198, 0, 1, 89, 148, 0, - 1, 60, 104, 0, 1, 33, 57, 0, 1, 12, 21, 0, 30, 181, 221, 0, 8, 141, 198, 0, - 1, 87, 145, 0, 1, 58, 100, 0, 1, 31, 55, 0, 1, 12, 20, 0, 32, 186, 224, 0, - 7, 142, 198, 0, 1, 86, 143, 0, 1, 58, 100, 0, 1, 31, 55, 0, 1, 12, 22, 0, - 57, 192, 227, 0, 20, 143, 204, 0, 3, 96, 154, 0, 1, 68, 112, 0, 1, 42, 69, 0, - 1, 19, 32, 0, 212, 35, 215, 0, 113, 47, 169, 0, 29, 48, 105, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 74, 129, 203, 0, 106, 120, 203, 0, 49, 107, 178, 0, - 19, 84, 144, 0, 4, 50, 84, 0, 1, 15, 25, 0, 71, 172, 217, 0, 44, 141, 209, 0, - 15, 102, 173, 0, 6, 76, 133, 0, 2, 51, 89, 0, 1, 24, 42, 0, 64, 185, 231, 0, - 31, 148, 216, 0, 8, 103, 175, 0, 3, 74, 131, 0, 1, 46, 81, 0, 1, 18, 30, 0, - 65, 196, 235, 0, 25, 157, 221, 0, 5, 105, 174, 0, 1, 67, 120, 0, 1, 38, 69, 0, - 1, 15, 30, 0, 65, 204, 238, 0, 30, 156, 224, 0, 7, 107, 177, 0, 2, 70, 124, 0, - 1, 42, 73, 0, 1, 18, 34, 0, 225, 86, 251, 0, 144, 104, 235, 0, 42, 99, 181, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 175, 239, 0, 112, 165, 229, 0, - 29, 136, 200, 0, 12, 103, 162, 0, 6, 77, 123, 0, 2, 53, 84, 0, 75, 183, 239, 0, - 30, 155, 221, 0, 3, 106, 171, 0, 1, 74, 128, 0, 1, 44, 76, 0, 1, 17, 28, 0, - 73, 185, 240, 0, 27, 159, 222, 0, 2, 107, 172, 0, 1, 75, 127, 0, 1, 42, 73, 0, - 1, 17, 29, 0, 62, 190, 238, 0, 21, 159, 222, 0, 2, 107, 172, 0, 1, 72, 122, 0, - 1, 40, 71, 0, 1, 18, 32, 0, 61, 199, 240, 0, 27, 161, 226, 0, 4, 113, 180, 0, - 1, 76, 129, 0, 1, 46, 80, 0, 1, 23, 41, 0, 7, 27, 153, 0, 5, 30, 95, 0, - 1, 16, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 50, 75, 127, 0, - 57, 75, 124, 0, 27, 67, 108, 0, 10, 54, 86, 0, 1, 33, 52, 0, 1, 12, 18, 0, - 43, 125, 151, 0, 26, 108, 148, 0, 7, 83, 122, 0, 2, 59, 89, 0, 1, 38, 60, 0, - 1, 17, 27, 0, 23, 144, 163, 0, 13, 112, 154, 0, 2, 75, 117, 0, 1, 50, 81, 0, - 1, 31, 51, 0, 1, 14, 23, 0, 18, 162, 185, 0, 6, 123, 171, 0, 1, 78, 125, 0, - 1, 51, 86, 0, 1, 31, 54, 0, 1, 14, 23, 0, 15, 199, 227, 0, 3, 150, 204, 0, - 1, 91, 146, 0, 1, 55, 95, 0, 1, 30, 53, 0, 1, 11, 20, 0, 19, 55, 240, 0, - 19, 59, 196, 0, 3, 52, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 41, 166, 207, 0, 104, 153, 199, 0, 31, 123, 181, 0, 14, 101, 152, 0, 5, 72, 106, 0, - 1, 36, 52, 0, 35, 176, 211, 0, 12, 131, 190, 0, 2, 88, 144, 0, 1, 60, 101, 0, - 1, 36, 60, 0, 1, 16, 28, 0, 28, 183, 213, 0, 8, 134, 191, 0, 1, 86, 142, 0, - 1, 56, 96, 0, 1, 30, 53, 0, 1, 12, 20, 0, 20, 190, 215, 0, 4, 135, 192, 0, - 1, 84, 139, 0, 1, 53, 91, 0, 1, 28, 49, 0, 1, 11, 20, 0, 13, 196, 216, 0, - 2, 137, 192, 0, 1, 86, 143, 0, 1, 57, 99, 0, 1, 32, 56, 0, 1, 13, 24, 0, - 211, 29, 217, 0, 96, 47, 156, 0, 22, 43, 87, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 78, 120, 193, 0, 111, 116, 186, 0, 46, 102, 164, 0, 15, 80, 128, 0, - 2, 49, 76, 0, 1, 18, 28, 0, 71, 161, 203, 0, 42, 132, 192, 0, 10, 98, 150, 0, - 3, 69, 109, 0, 1, 44, 70, 0, 1, 18, 29, 0, 57, 186, 211, 0, 30, 140, 196, 0, - 4, 93, 146, 0, 1, 62, 102, 0, 1, 38, 65, 0, 1, 16, 27, 0, 47, 199, 217, 0, - 14, 145, 196, 0, 1, 88, 142, 0, 1, 57, 98, 0, 1, 36, 62, 0, 1, 15, 26, 0, - 26, 219, 229, 0, 5, 155, 207, 0, 1, 94, 151, 0, 1, 60, 104, 0, 1, 36, 62, 0, - 1, 16, 28, 0, 233, 29, 248, 0, 146, 47, 220, 0, 43, 52, 140, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 100, 163, 232, 0, 179, 161, 222, 0, 63, 142, 204, 0, - 37, 113, 174, 0, 26, 89, 137, 0, 18, 68, 97, 0, 85, 181, 230, 0, 32, 146, 209, 0, - 7, 100, 164, 0, 3, 71, 121, 0, 1, 45, 77, 0, 1, 18, 30, 0, 65, 187, 230, 0, - 20, 148, 207, 0, 2, 97, 159, 0, 1, 68, 116, 0, 1, 40, 70, 0, 1, 14, 29, 0, - 40, 194, 227, 0, 8, 147, 204, 0, 1, 94, 155, 0, 1, 65, 112, 0, 1, 39, 66, 0, - 1, 14, 26, 0, 16, 208, 228, 0, 3, 151, 207, 0, 1, 98, 160, 0, 1, 67, 117, 0, - 1, 41, 74, 0, 1, 17, 31, 0, 17, 38, 140, 0, 7, 34, 80, 0, 1, 17, 29, 0, - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 37, 75, 128, 0, 41, 76, 128, 0, - 26, 66, 116, 0, 12, 52, 94, 0, 2, 32, 55, 0, 1, 10, 16, 0, 50, 127, 154, 0, - 37, 109, 152, 0, 16, 82, 121, 0, 5, 59, 85, 0, 1, 35, 54, 0, 1, 13, 20, 0, - 40, 142, 167, 0, 17, 110, 157, 0, 2, 71, 112, 0, 1, 44, 72, 0, 1, 27, 45, 0, - 1, 11, 17, 0, 30, 175, 188, 0, 9, 124, 169, 0, 1, 74, 116, 0, 1, 48, 78, 0, - 1, 30, 49, 0, 1, 11, 18, 0, 10, 222, 223, 0, 2, 150, 194, 0, 1, 83, 128, 0, - 1, 48, 79, 0, 1, 27, 45, 0, 1, 11, 17, 0, 36, 41, 235, 0, 29, 36, 193, 0, - 10, 27, 111, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 85, 165, 222, 0, - 177, 162, 215, 0, 110, 135, 195, 0, 57, 113, 168, 0, 23, 83, 120, 0, 10, 49, 61, 0, - 85, 190, 223, 0, 36, 139, 200, 0, 5, 90, 146, 0, 1, 60, 103, 0, 1, 38, 65, 0, - 1, 18, 30, 0, 72, 202, 223, 0, 23, 141, 199, 0, 2, 86, 140, 0, 1, 56, 97, 0, - 1, 36, 61, 0, 1, 16, 27, 0, 55, 218, 225, 0, 13, 145, 200, 0, 1, 86, 141, 0, - 1, 57, 99, 0, 1, 35, 61, 0, 1, 13, 22, 0, 15, 235, 212, 0, 1, 132, 184, 0, - 1, 84, 139, 0, 1, 57, 97, 0, 1, 34, 56, 0, 1, 14, 23, 0, 181, 21, 201, 0, - 61, 37, 123, 0, 10, 38, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 47, 106, 172, 0, 95, 104, 173, 0, 42, 93, 159, 0, 18, 77, 131, 0, 4, 50, 81, 0, - 1, 17, 23, 0, 62, 147, 199, 0, 44, 130, 189, 0, 28, 102, 154, 0, 18, 75, 115, 0, - 2, 44, 65, 0, 1, 12, 19, 0, 55, 153, 210, 0, 24, 130, 194, 0, 3, 93, 146, 0, - 1, 61, 97, 0, 1, 31, 50, 0, 1, 10, 16, 0, 49, 186, 223, 0, 17, 148, 204, 0, - 1, 96, 142, 0, 1, 53, 83, 0, 1, 26, 44, 0, 1, 11, 17, 0, 13, 217, 212, 0, - 2, 136, 180, 0, 1, 78, 124, 0, 1, 50, 83, 0, 1, 29, 49, 0, 1, 14, 23, 0, - 197, 13, 247, 0, 82, 17, 222, 0, 25, 17, 162, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0, 0, 126, 186, 247, 0, 234, 191, 243, 0, 176, 177, 234, 0, 104, 158, 220, 0, - 66, 128, 186, 0, 55, 90, 137, 0, 111, 197, 242, 0, 46, 158, 219, 0, 9, 104, 171, 0, - 2, 65, 125, 0, 1, 44, 80, 0, 1, 17, 91, 0, 104, 208, 245, 0, 39, 168, 224, 0, - 3, 109, 162, 0, 1, 79, 124, 0, 1, 50, 102, 0, 1, 43, 102, 0, 84, 220, 246, 0, - 31, 177, 231, 0, 2, 115, 180, 0, 1, 79, 134, 0, 1, 55, 77, 0, 1, 60, 79, 0, - 43, 243, 240, 0, 8, 180, 217, 0, 1, 115, 166, 0, 1, 84, 121, 0, 1, 51, 67, 0, - 1, 16, 6, 0, + 195, 29, 183, 84, 49, 136, 8, 42, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 31, 107, 169, 35, 99, 159, 17, 82, 140, 8, 66, 114, 2, 44, 76, 1, 19, 32, + 40, 132, 201, 29, 114, 187, 13, 91, 157, 7, 75, 127, 3, 58, 95, 1, 28, 47, + 69, 142, 221, 42, 122, 201, 15, 91, 159, 6, 67, 121, 1, 42, 77, 1, 17, 31, + 102, 148, 228, 67, 117, 204, 17, 82, 154, 6, 59, 114, 2, 39, 75, 1, 15, 29, + 156, 57, 233, 119, 57, 212, 58, 48, 163, 29, 40, 124, 12, 30, 81, 3, 12, 31, + 191, 107, 226, 124, 117, 204, 25, 99, 155, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 29, 148, 210, 37, 126, 194, 8, 93, 157, 2, 68, 118, 1, 39, 69, 1, 17, 33, + 41, 151, 213, 27, 123, 193, 3, 82, 144, 1, 58, 105, 1, 32, 60, 1, 13, 26, + 59, 159, 220, 23, 126, 198, 4, 88, 151, 1, 66, 114, 1, 38, 71, 1, 18, 34, + 114, 136, 232, 51, 114, 207, 11, 83, 155, 3, 56, 105, 1, 33, 65, 1, 17, 34, + 149, 65, 234, 121, 57, 215, 61, 49, 166, 28, 36, 114, 12, 25, 76, 3, 16, 42, + 214, 49, 220, 132, 63, 188, 42, 65, 137, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 85, 137, 221, 104, 131, 216, 49, 111, 192, 21, 87, 155, 2, 49, 87, 1, 16, 28, + 89, 163, 230, 90, 137, 220, 29, 100, 183, 10, 70, 135, 2, 42, 81, 1, 17, 33, + 108, 167, 237, 55, 133, 222, 15, 97, 179, 4, 72, 135, 1, 45, 85, 1, 19, 38, + 124, 146, 240, 66, 124, 224, 17, 88, 175, 4, 58, 122, 1, 36, 75, 1, 18, 37, + 141, 79, 241, 126, 70, 227, 66, 58, 182, 30, 44, 136, 12, 34, 96, 2, 20, 47, + 229, 99, 249, 143, 111, 235, 46, 109, 192, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 82, 158, 236, 94, 146, 224, 25, 117, 191, 9, 87, 149, 3, 56, 99, 1, 33, 57, + 83, 167, 237, 68, 145, 222, 10, 103, 177, 2, 72, 131, 1, 41, 79, 1, 20, 39, + 99, 167, 239, 47, 141, 224, 10, 104, 178, 2, 73, 133, 1, 44, 85, 1, 22, 47, + 127, 145, 243, 71, 129, 228, 17, 93, 177, 3, 61, 124, 1, 41, 84, 1, 21, 52, + 157, 78, 244, 140, 72, 231, 69, 58, 184, 31, 44, 137, 14, 38, 105, 8, 23, 61, + 125, 34, 187, 52, 41, 133, 6, 31, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 37, 109, 153, 51, 102, 147, 23, 87, 128, 8, 67, 101, 1, 41, 63, 1, 19, 29, + 31, 154, 185, 17, 127, 175, 6, 96, 145, 2, 73, 114, 1, 51, 82, 1, 28, 45, + 23, 163, 200, 10, 131, 185, 2, 93, 148, 1, 67, 111, 1, 41, 69, 1, 14, 24, + 29, 176, 217, 12, 145, 201, 3, 101, 156, 1, 69, 111, 1, 39, 63, 1, 14, 23, + 57, 192, 233, 25, 154, 215, 6, 109, 167, 3, 78, 118, 1, 48, 69, 1, 21, 29, + 202, 105, 245, 108, 106, 216, 18, 90, 144, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 33, 172, 219, 64, 149, 206, 14, 117, 177, 5, 90, 141, 2, 61, 95, 1, 37, 57, + 33, 179, 220, 11, 140, 198, 1, 89, 148, 1, 60, 104, 1, 33, 57, 1, 12, 21, + 30, 181, 221, 8, 141, 198, 1, 87, 145, 1, 58, 100, 1, 31, 55, 1, 12, 20, + 32, 186, 224, 7, 142, 198, 1, 86, 143, 1, 58, 100, 1, 31, 55, 1, 12, 22, + 57, 192, 227, 20, 143, 204, 3, 96, 154, 1, 68, 112, 1, 42, 69, 1, 19, 32, + 212, 35, 215, 113, 47, 169, 29, 48, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 74, 129, 203, 106, 120, 203, 49, 107, 178, 19, 84, 144, 4, 50, 84, 1, 15, 25, + 71, 172, 217, 44, 141, 209, 15, 102, 173, 6, 76, 133, 2, 51, 89, 1, 24, 42, + 64, 185, 231, 31, 148, 216, 8, 103, 175, 3, 74, 131, 1, 46, 81, 1, 18, 30, + 65, 196, 235, 25, 157, 221, 5, 105, 174, 1, 67, 120, 1, 38, 69, 1, 15, 30, + 65, 204, 238, 30, 156, 224, 7, 107, 177, 2, 70, 124, 1, 42, 73, 1, 18, 34, + 225, 86, 251, 144, 104, 235, 42, 99, 181, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 85, 175, 239, 112, 165, 229, 29, 136, 200, 12, 103, 162, 6, 77, 123, 2, 53, 84, + 75, 183, 239, 30, 155, 221, 3, 106, 171, 1, 74, 128, 1, 44, 76, 1, 17, 28, + 73, 185, 240, 27, 159, 222, 2, 107, 172, 1, 75, 127, 1, 42, 73, 1, 17, 29, + 62, 190, 238, 21, 159, 222, 2, 107, 172, 1, 72, 122, 1, 40, 71, 1, 18, 32, + 61, 199, 240, 27, 161, 226, 4, 113, 180, 1, 76, 129, 1, 46, 80, 1, 23, 41, + 7, 27, 153, 5, 30, 95, 1, 16, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 50, 75, 127, 57, 75, 124, 27, 67, 108, 10, 54, 86, 1, 33, 52, 1, 12, 18, + 43, 125, 151, 26, 108, 148, 7, 83, 122, 2, 59, 89, 1, 38, 60, 1, 17, 27, + 23, 144, 163, 13, 112, 154, 2, 75, 117, 1, 50, 81, 1, 31, 51, 1, 14, 23, + 18, 162, 185, 6, 123, 171, 1, 78, 125, 1, 51, 86, 1, 31, 54, 1, 14, 23, + 15, 199, 227, 3, 150, 204, 1, 91, 146, 1, 55, 95, 1, 30, 53, 1, 11, 20, + 19, 55, 240, 19, 59, 196, 3, 52, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 41, 166, 207, 104, 153, 199, 31, 123, 181, 14, 101, 152, 5, 72, 106, 1, 36, 52, + 35, 176, 211, 12, 131, 190, 2, 88, 144, 1, 60, 101, 1, 36, 60, 1, 16, 28, + 28, 183, 213, 8, 134, 191, 1, 86, 142, 1, 56, 96, 1, 30, 53, 1, 12, 20, + 20, 190, 215, 4, 135, 192, 1, 84, 139, 1, 53, 91, 1, 28, 49, 1, 11, 20, + 13, 196, 216, 2, 137, 192, 1, 86, 143, 1, 57, 99, 1, 32, 56, 1, 13, 24, + 211, 29, 217, 96, 47, 156, 22, 43, 87, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 78, 120, 193, 111, 116, 186, 46, 102, 164, 15, 80, 128, 2, 49, 76, 1, 18, 28, + 71, 161, 203, 42, 132, 192, 10, 98, 150, 3, 69, 109, 1, 44, 70, 1, 18, 29, + 57, 186, 211, 30, 140, 196, 4, 93, 146, 1, 62, 102, 1, 38, 65, 1, 16, 27, + 47, 199, 217, 14, 145, 196, 1, 88, 142, 1, 57, 98, 1, 36, 62, 1, 15, 26, + 26, 219, 229, 5, 155, 207, 1, 94, 151, 1, 60, 104, 1, 36, 62, 1, 16, 28, + 233, 29, 248, 146, 47, 220, 43, 52, 140, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 100, 163, 232, 179, 161, 222, 63, 142, 204, 37, 113, 174, 26, 89, 137, 18, 68, 97, + 85, 181, 230, 32, 146, 209, 7, 100, 164, 3, 71, 121, 1, 45, 77, 1, 18, 30, + 65, 187, 230, 20, 148, 207, 2, 97, 159, 1, 68, 116, 1, 40, 70, 1, 14, 29, + 40, 194, 227, 8, 147, 204, 1, 94, 155, 1, 65, 112, 1, 39, 66, 1, 14, 26, + 16, 208, 228, 3, 151, 207, 1, 98, 160, 1, 67, 117, 1, 41, 74, 1, 17, 31, + 17, 38, 140, 7, 34, 80, 1, 17, 29, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 37, 75, 128, 41, 76, 128, 26, 66, 116, 12, 52, 94, 2, 32, 55, 1, 10, 16, + 50, 127, 154, 37, 109, 152, 16, 82, 121, 5, 59, 85, 1, 35, 54, 1, 13, 20, + 40, 142, 167, 17, 110, 157, 2, 71, 112, 1, 44, 72, 1, 27, 45, 1, 11, 17, + 30, 175, 188, 9, 124, 169, 1, 74, 116, 1, 48, 78, 1, 30, 49, 1, 11, 18, + 10, 222, 223, 2, 150, 194, 1, 83, 128, 1, 48, 79, 1, 27, 45, 1, 11, 17, + 36, 41, 235, 29, 36, 193, 10, 27, 111, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 85, 165, 222, 177, 162, 215, 110, 135, 195, 57, 113, 168, 23, 83, 120, 10, 49, 61, + 85, 190, 223, 36, 139, 200, 5, 90, 146, 1, 60, 103, 1, 38, 65, 1, 18, 30, + 72, 202, 223, 23, 141, 199, 2, 86, 140, 1, 56, 97, 1, 36, 61, 1, 16, 27, + 55, 218, 225, 13, 145, 200, 1, 86, 141, 1, 57, 99, 1, 35, 61, 1, 13, 22, + 15, 235, 212, 1, 132, 184, 1, 84, 139, 1, 57, 97, 1, 34, 56, 1, 14, 23, + 181, 21, 201, 61, 37, 123, 10, 38, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 47, 106, 172, 95, 104, 173, 42, 93, 159, 18, 77, 131, 4, 50, 81, 1, 17, 23, + 62, 147, 199, 44, 130, 189, 28, 102, 154, 18, 75, 115, 2, 44, 65, 1, 12, 19, + 55, 153, 210, 24, 130, 194, 3, 93, 146, 1, 61, 97, 1, 31, 50, 1, 10, 16, + 49, 186, 223, 17, 148, 204, 1, 96, 142, 1, 53, 83, 1, 26, 44, 1, 11, 17, + 13, 217, 212, 2, 136, 180, 1, 78, 124, 1, 50, 83, 1, 29, 49, 1, 14, 23, + 197, 13, 247, 82, 17, 222, 25, 17, 162, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 126, 186, 247, 234, 191, 243, 176, 177, 234, 104, 158, 220, 66, 128, 186, 55, 90, 137, + 111, 197, 242, 46, 158, 219, 9, 104, 171, 2, 65, 125, 1, 44, 80, 1, 17, 91, + 104, 208, 245, 39, 168, 224, 3, 109, 162, 1, 79, 124, 1, 50, 102, 1, 43, 102, + 84, 220, 246, 31, 177, 231, 2, 115, 180, 1, 79, 134, 1, 55, 77, 1, 60, 79, + 43, 243, 240, 8, 180, 217, 1, 115, 166, 1, 84, 121, 1, 51, 67, 1, 16, 6, }, .switchable_interp_prob{235, 162, 36, 255, 34, 3, 149, 144}, .inter_mode_prob{ @@ -322,39 +302,23 @@ bool VP9::WriteLessThan(VpxRangeEncoder& writer, s32 value, s32 test) { } void VP9::WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, - const std::array& new_prob, - const std::array& old_prob) { - // Note: There's 1 byte added on each packet for alignment, - // this byte is ignored when doing updates. - constexpr s32 block_bytes = 2 * 2 * 6 * 6 * 4; - - const auto needs_update = [&](s32 base_index) -> bool { - s32 index = base_index; - for (s32 i = 0; i < 2; i++) { - for (s32 j = 0; j < 2; j++) { - for (s32 k = 0; k < 6; k++) { - for (s32 l = 0; l < 6; l++) { - if (new_prob[index + 0] != old_prob[index + 0] || - new_prob[index + 1] != old_prob[index + 1] || - new_prob[index + 2] != old_prob[index + 2]) { - return true; - } - - index += 4; - } - } - } - } - return false; + const std::array& new_prob, + const std::array& old_prob) { + constexpr u32 block_bytes = 2 * 2 * 6 * 6 * 3; + + const auto needs_update = [&](u32 base_index) -> bool { + return !std::equal(new_prob.begin() + base_index, + new_prob.begin() + base_index + block_bytes, + old_prob.begin() + base_index); }; - for (s32 block_index = 0; block_index < 4; block_index++) { - const s32 base_index = block_index * block_bytes; + for (u32 block_index = 0; block_index < 4; block_index++) { + const u32 base_index = block_index * block_bytes; const bool update = needs_update(base_index); writer.Write(update); if (update) { - s32 index = base_index; + u32 index = base_index; for (s32 i = 0; i < 2; i++) { for (s32 j = 0; j < 2; j++) { for (s32 k = 0; k < 6; k++) { @@ -367,14 +331,13 @@ void VP9::WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, WriteProbabilityUpdate(writer, new_prob[index + 2], old_prob[index + 2]); } - index += 4; + index += 3; } } } } } - - if (block_index == tx_mode) { + if (block_index == static_cast(tx_mode)) { break; } } @@ -392,7 +355,7 @@ void VP9::WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_ Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) { PictureInfo picture_info{}; gpu.MemoryManager().ReadBlock(state.picture_info_offset, &picture_info, sizeof(PictureInfo)); - Vp9PictureInfo vp9_info = picture_info.Convert(); + Vp9PictureInfo vp9_info = std::move(picture_info.Convert()); InsertEntropy(state.vp9_entropy_probs_offset, vp9_info.entropy); @@ -414,8 +377,7 @@ Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) Vp9FrameContainer frame{}; { gpu.SyncGuestHost(); - frame.info = GetVp9PictureInfo(state); - + frame.info = std::move(GetVp9PictureInfo(state)); frame.bit_stream.resize(frame.info.bitstream_size); gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, frame.bit_stream.data(), frame.info.bitstream_size); @@ -423,37 +385,37 @@ Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) // Buffer two frames, saving the last show frame info if (!next_next_frame.bit_stream.empty()) { Vp9FrameContainer temp{ - .info = frame.info, - .bit_stream = frame.bit_stream, + .info = std::move(frame.info), + .bit_stream = std::move(frame.bit_stream), }; next_next_frame.info.show_frame = frame.info.last_frame_shown; - frame.info = next_next_frame.info; - frame.bit_stream = next_next_frame.bit_stream; + frame.info = std::move(next_next_frame.info); + frame.bit_stream = std::move(next_next_frame.bit_stream); next_next_frame = std::move(temp); if (!next_frame.bit_stream.empty()) { Vp9FrameContainer temp2{ - .info = frame.info, - .bit_stream = frame.bit_stream, + .info = std::move(frame.info), + .bit_stream = std::move(frame.bit_stream), }; next_frame.info.show_frame = frame.info.last_frame_shown; - frame.info = next_frame.info; - frame.bit_stream = next_frame.bit_stream; + frame.info = std::move(next_frame.info); + frame.bit_stream = std::move(next_frame.bit_stream); next_frame = std::move(temp2); } else { - next_frame.info = frame.info; - next_frame.bit_stream = frame.bit_stream; + next_frame.info = std::move(frame.info); + next_frame.bit_stream = std::move(frame.bit_stream); } } else { - next_next_frame.info = frame.info; - next_next_frame.bit_stream = frame.bit_stream; + next_next_frame.info = std::move(frame.info); + next_next_frame.bit_stream = std::move(frame.bit_stream); } return frame; } std::vector VP9::ComposeCompressedHeader() { VpxRangeEncoder writer{}; - + const bool update_probs = current_frame_info.show_frame && !current_frame_info.is_key_frame; if (!current_frame_info.lossless) { if (static_cast(current_frame_info.transform_mode) >= 3) { writer.Write(3, 2); @@ -471,7 +433,7 @@ std::vector VP9::ComposeCompressedHeader() { prev_frame_probs.tx_16x16_prob); WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_32x32_prob, prev_frame_probs.tx_32x32_prob); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + if (update_probs) { prev_frame_probs.tx_8x8_prob = current_frame_info.entropy.tx_8x8_prob; prev_frame_probs.tx_16x16_prob = current_frame_info.entropy.tx_16x16_prob; prev_frame_probs.tx_32x32_prob = current_frame_info.entropy.tx_32x32_prob; @@ -484,7 +446,7 @@ std::vector VP9::ComposeCompressedHeader() { WriteProbabilityUpdate(writer, current_frame_info.entropy.skip_probs, prev_frame_probs.skip_probs); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + if (update_probs) { prev_frame_probs.coef_probs = current_frame_info.entropy.coef_probs; prev_frame_probs.skip_probs = current_frame_info.entropy.skip_probs; } @@ -493,15 +455,12 @@ std::vector VP9::ComposeCompressedHeader() { // read_inter_probs() in the spec WriteProbabilityUpdateAligned4(writer, current_frame_info.entropy.inter_mode_prob, prev_frame_probs.inter_mode_prob); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { - prev_frame_probs.inter_mode_prob = current_frame_info.entropy.inter_mode_prob; - } if (current_frame_info.interp_filter == 4) { // read_interp_filter_probs() in the spec WriteProbabilityUpdate(writer, current_frame_info.entropy.switchable_interp_prob, prev_frame_probs.switchable_interp_prob); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + if (update_probs) { prev_frame_probs.switchable_interp_prob = current_frame_info.entropy.switchable_interp_prob; } @@ -510,9 +469,7 @@ std::vector VP9::ComposeCompressedHeader() { // read_is_inter_probs() in the spec WriteProbabilityUpdate(writer, current_frame_info.entropy.intra_inter_prob, prev_frame_probs.intra_inter_prob); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { - prev_frame_probs.intra_inter_prob = current_frame_info.entropy.intra_inter_prob; - } + // frame_reference_mode() in the spec if ((current_frame_info.ref_frame_sign_bias[1] & 1) != (current_frame_info.ref_frame_sign_bias[2] & 1) || @@ -530,7 +487,7 @@ std::vector VP9::ComposeCompressedHeader() { if (current_frame_info.reference_mode == 2) { WriteProbabilityUpdate(writer, current_frame_info.entropy.comp_inter_prob, prev_frame_probs.comp_inter_prob); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + if (update_probs) { prev_frame_probs.comp_inter_prob = current_frame_info.entropy.comp_inter_prob; } } @@ -538,7 +495,7 @@ std::vector VP9::ComposeCompressedHeader() { if (current_frame_info.reference_mode != 1) { WriteProbabilityUpdate(writer, current_frame_info.entropy.single_ref_prob, prev_frame_probs.single_ref_prob); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + if (update_probs) { prev_frame_probs.single_ref_prob = current_frame_info.entropy.single_ref_prob; } } @@ -546,7 +503,7 @@ std::vector VP9::ComposeCompressedHeader() { if (current_frame_info.reference_mode != 0) { WriteProbabilityUpdate(writer, current_frame_info.entropy.comp_ref_prob, prev_frame_probs.comp_ref_prob); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + if (update_probs) { prev_frame_probs.comp_ref_prob = current_frame_info.entropy.comp_ref_prob; } } @@ -557,42 +514,37 @@ std::vector VP9::ComposeCompressedHeader() { WriteProbabilityUpdate(writer, current_frame_info.entropy.y_mode_prob[index], prev_frame_probs.y_mode_prob[index]); } - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { - prev_frame_probs.y_mode_prob = current_frame_info.entropy.y_mode_prob; - } + // read_partition_probs WriteProbabilityUpdateAligned4(writer, current_frame_info.entropy.partition_prob, prev_frame_probs.partition_prob); - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { - prev_frame_probs.partition_prob = current_frame_info.entropy.partition_prob; - } // mv_probs for (s32 i = 0; i < 3; i++) { WriteMvProbabilityUpdate(writer, current_frame_info.entropy.joints[i], prev_frame_probs.joints[i]); } - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + if (update_probs) { + prev_frame_probs.inter_mode_prob = current_frame_info.entropy.inter_mode_prob; + prev_frame_probs.intra_inter_prob = current_frame_info.entropy.intra_inter_prob; + prev_frame_probs.y_mode_prob = current_frame_info.entropy.y_mode_prob; + prev_frame_probs.partition_prob = current_frame_info.entropy.partition_prob; prev_frame_probs.joints = current_frame_info.entropy.joints; } for (s32 i = 0; i < 2; i++) { WriteMvProbabilityUpdate(writer, current_frame_info.entropy.sign[i], prev_frame_probs.sign[i]); - for (s32 j = 0; j < 10; j++) { const int index = i * 10 + j; - WriteMvProbabilityUpdate(writer, current_frame_info.entropy.classes[index], prev_frame_probs.classes[index]); } - WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0[i], prev_frame_probs.class_0[i]); for (s32 j = 0; j < 10; j++) { const int index = i * 10 + j; - WriteMvProbabilityUpdate(writer, current_frame_info.entropy.prob_bits[index], prev_frame_probs.prob_bits[index]); } @@ -602,7 +554,6 @@ std::vector VP9::ComposeCompressedHeader() { for (s32 j = 0; j < 2; j++) { for (s32 k = 0; k < 3; k++) { const int index = i * 2 * 3 + j * 3 + k; - WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0_fr[index], prev_frame_probs.class_0_fr[index]); } @@ -610,7 +561,6 @@ std::vector VP9::ComposeCompressedHeader() { for (s32 j = 0; j < 3; j++) { const int index = i * 3 + j; - WriteMvProbabilityUpdate(writer, current_frame_info.entropy.fr[index], prev_frame_probs.fr[index]); } @@ -626,7 +576,7 @@ std::vector VP9::ComposeCompressedHeader() { } // save previous probs - if (current_frame_info.show_frame && !current_frame_info.is_key_frame) { + if (update_probs) { prev_frame_probs.sign = current_frame_info.entropy.sign; prev_frame_probs.classes = current_frame_info.entropy.classes; prev_frame_probs.class_0 = current_frame_info.entropy.class_0; @@ -637,7 +587,6 @@ std::vector VP9::ComposeCompressedHeader() { prev_frame_probs.high_precision = current_frame_info.entropy.high_precision; } } - writer.End(); return writer.GetBuffer(); } @@ -854,11 +803,11 @@ VpxBitStreamWriter VP9::ComposeUncompressedHeader() { return uncomp_writer; } -const std::vector& VP9::ComposeFrameHeader(NvdecCommon::NvdecRegisters& state) { +const std::vector& VP9::ComposeFrameHeader(const NvdecCommon::NvdecRegisters& state) { std::vector bitstream; { - Vp9FrameContainer curr_frame = GetCurrentFrame(state); - current_frame_info = curr_frame.info; + Vp9FrameContainer curr_frame = std::move(GetCurrentFrame(state)); + current_frame_info = std::move(curr_frame.info); bitstream = std::move(curr_frame.bit_stream); } diff --git a/src/video_core/command_classes/codecs/vp9.h b/src/video_core/command_classes/codecs/vp9.h index e2504512c7..9ebbbf59e5 100644 --- a/src/video_core/command_classes/codecs/vp9.h +++ b/src/video_core/command_classes/codecs/vp9.h @@ -119,7 +119,8 @@ public: /// Composes the VP9 frame from the GPU state information. Based on the official VP9 spec /// documentation - [[nodiscard]] const std::vector& ComposeFrameHeader(NvdecCommon::NvdecRegisters& state); + [[nodiscard]] const std::vector& ComposeFrameHeader( + const NvdecCommon::NvdecRegisters& state); /// Returns true if the most recent frame was a hidden frame. [[nodiscard]] bool WasFrameHidden() const { @@ -147,8 +148,8 @@ private: /// Writes probability updates for the Coef probabilities void WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, - const std::array& new_prob, - const std::array& old_prob); + const std::array& new_prob, + const std::array& old_prob); /// Write probabilities for 4-byte aligned structures template diff --git a/src/video_core/command_classes/codecs/vp9_types.h b/src/video_core/command_classes/codecs/vp9_types.h index 4f0b05d225..5ca944f2ac 100644 --- a/src/video_core/command_classes/codecs/vp9_types.h +++ b/src/video_core/command_classes/codecs/vp9_types.h @@ -31,62 +31,6 @@ enum FrameFlags : u32 { IntraOnly = 1 << 5, }; -enum class MvJointType { - MvJointZero = 0, /* Zero vector */ - MvJointHnzvz = 1, /* Vert zero, hor nonzero */ - MvJointHzvnz = 2, /* Hor zero, vert nonzero */ - MvJointHnzvnz = 3, /* Both components nonzero */ -}; -enum class MvClassType { - MvClass0 = 0, /* (0, 2] integer pel */ - MvClass1 = 1, /* (2, 4] integer pel */ - MvClass2 = 2, /* (4, 8] integer pel */ - MvClass3 = 3, /* (8, 16] integer pel */ - MvClass4 = 4, /* (16, 32] integer pel */ - MvClass5 = 5, /* (32, 64] integer pel */ - MvClass6 = 6, /* (64, 128] integer pel */ - MvClass7 = 7, /* (128, 256] integer pel */ - MvClass8 = 8, /* (256, 512] integer pel */ - MvClass9 = 9, /* (512, 1024] integer pel */ - MvClass10 = 10, /* (1024,2048] integer pel */ -}; - -enum class BlockSize { - Block4x4 = 0, - Block4x8 = 1, - Block8x4 = 2, - Block8x8 = 3, - Block8x16 = 4, - Block16x8 = 5, - Block16x16 = 6, - Block16x32 = 7, - Block32x16 = 8, - Block32x32 = 9, - Block32x64 = 10, - Block64x32 = 11, - Block64x64 = 12, - BlockSizes = 13, - BlockInvalid = BlockSizes -}; - -enum class PredictionMode { - DcPred = 0, // Average of above and left pixels - VPred = 1, // Vertical - HPred = 2, // Horizontal - D45Pred = 3, // Directional 45 deg = round(arctan(1 / 1) * 180 / pi) - D135Pred = 4, // Directional 135 deg = 180 - 45 - D117Pred = 5, // Directional 117 deg = 180 - 63 - D153Pred = 6, // Directional 153 deg = 180 - 27 - D207Pred = 7, // Directional 207 deg = 180 + 27 - D63Pred = 8, // Directional 63 deg = round(arctan(2 / 1) * 180 / pi) - TmPred = 9, // True-motion - NearestMv = 10, - NearMv = 11, - ZeroMv = 12, - NewMv = 13, - MbModeCount = 14 -}; - enum class TxSize { Tx4x4 = 0, // 4x4 transform Tx8x8 = 1, // 8x8 transform @@ -104,13 +48,6 @@ enum class TxMode { TxModes = 5 }; -enum class reference_mode { - SingleReference = 0, - CompoundReference = 1, - ReferenceModeSelect = 2, - ReferenceModes = 3 -}; - struct Segmentation { u8 enabled{}; u8 update_map{}; @@ -131,7 +68,7 @@ static_assert(sizeof(LoopFilter) == 0x7, "LoopFilter is an invalid size"); struct Vp9EntropyProbs { std::array y_mode_prob{}; std::array partition_prob{}; - std::array coef_probs{}; + std::array coef_probs{}; std::array switchable_interp_prob{}; std::array inter_mode_prob{}; std::array intra_inter_prob{}; @@ -152,7 +89,7 @@ struct Vp9EntropyProbs { std::array class_0_hp{}; std::array high_precision{}; }; -static_assert(sizeof(Vp9EntropyProbs) == 0x9F4, "Vp9EntropyProbs is an invalid size"); +static_assert(sizeof(Vp9EntropyProbs) == 0x7B4, "Vp9EntropyProbs is an invalid size"); struct Vp9PictureInfo { bool is_key_frame{}; @@ -278,44 +215,41 @@ static_assert(sizeof(PictureInfo) == 0x100, "PictureInfo is an invalid size"); struct EntropyProbs { INSERT_PADDING_BYTES(1024); - std::array, 7> inter_mode_prob{}; + std::array inter_mode_prob{}; std::array intra_inter_prob{}; INSERT_PADDING_BYTES(80); - std::array, 2> tx_8x8_prob{}; - std::array, 2> tx_16x16_prob{}; - std::array, 2> tx_32x32_prob{}; + std::array tx_8x8_prob{}; + std::array tx_16x16_prob{}; + std::array tx_32x32_prob{}; std::array y_mode_prob_e8{}; std::array, 4> y_mode_prob_e0e7{}; INSERT_PADDING_BYTES(64); - std::array, 16> partition_prob{}; + std::array partition_prob{}; INSERT_PADDING_BYTES(10); - std::array, 4> switchable_interp_prob{}; + std::array switchable_interp_prob{}; std::array comp_inter_prob{}; - std::array skip_probs{}; + std::array skip_probs{}; + INSERT_PADDING_BYTES(1); std::array joints{}; std::array sign{}; - std::array, 2> class_0{}; - std::array, 2> fr{}; + std::array class_0{}; + std::array fr{}; std::array class_0_hp{}; std::array high_precision{}; - std::array, 2> classes{}; - std::array, 2>, 2> class_0_fr{}; - std::array, 2> pred_bits{}; - std::array, 5> single_ref_prob{}; + std::array classes{}; + std::array class_0_fr{}; + std::array pred_bits{}; + std::array single_ref_prob{}; std::array comp_ref_prob{}; INSERT_PADDING_BYTES(17); - std::array, 6>, 6>, 2>, 2>, 4> - coef_probs{}; + std::array coef_probs{}; void Convert(Vp9EntropyProbs& fc) { - std::memcpy(fc.inter_mode_prob.data(), inter_mode_prob.data(), fc.inter_mode_prob.size()); - - std::memcpy(fc.intra_inter_prob.data(), intra_inter_prob.data(), - fc.intra_inter_prob.size()); - - std::memcpy(fc.tx_8x8_prob.data(), tx_8x8_prob.data(), fc.tx_8x8_prob.size()); - std::memcpy(fc.tx_16x16_prob.data(), tx_16x16_prob.data(), fc.tx_16x16_prob.size()); - std::memcpy(fc.tx_32x32_prob.data(), tx_32x32_prob.data(), fc.tx_32x32_prob.size()); + fc.inter_mode_prob = std::move(inter_mode_prob); + fc.intra_inter_prob = std::move(intra_inter_prob); + fc.tx_8x8_prob = std::move(tx_8x8_prob); + fc.tx_16x16_prob = std::move(tx_16x16_prob); + fc.tx_32x32_prob = std::move(tx_32x32_prob); for (s32 i = 0; i < 4; i++) { for (s32 j = 0; j < 9; j++) { @@ -323,27 +257,29 @@ struct EntropyProbs { } } - std::memcpy(fc.partition_prob.data(), partition_prob.data(), fc.partition_prob.size()); - - std::memcpy(fc.switchable_interp_prob.data(), switchable_interp_prob.data(), - fc.switchable_interp_prob.size()); - std::memcpy(fc.comp_inter_prob.data(), comp_inter_prob.data(), fc.comp_inter_prob.size()); - std::memcpy(fc.skip_probs.data(), skip_probs.data(), fc.skip_probs.size()); - - std::memcpy(fc.joints.data(), joints.data(), fc.joints.size()); - - std::memcpy(fc.sign.data(), sign.data(), fc.sign.size()); - std::memcpy(fc.class_0.data(), class_0.data(), fc.class_0.size()); - std::memcpy(fc.fr.data(), fr.data(), fc.fr.size()); - std::memcpy(fc.class_0_hp.data(), class_0_hp.data(), fc.class_0_hp.size()); - std::memcpy(fc.high_precision.data(), high_precision.data(), fc.high_precision.size()); - std::memcpy(fc.classes.data(), classes.data(), fc.classes.size()); - std::memcpy(fc.class_0_fr.data(), class_0_fr.data(), fc.class_0_fr.size()); - std::memcpy(fc.prob_bits.data(), pred_bits.data(), fc.prob_bits.size()); - std::memcpy(fc.single_ref_prob.data(), single_ref_prob.data(), fc.single_ref_prob.size()); - std::memcpy(fc.comp_ref_prob.data(), comp_ref_prob.data(), fc.comp_ref_prob.size()); - - std::memcpy(fc.coef_probs.data(), coef_probs.data(), fc.coef_probs.size()); + fc.partition_prob = std::move(partition_prob); + fc.switchable_interp_prob = std::move(switchable_interp_prob); + fc.comp_inter_prob = std::move(comp_inter_prob); + fc.skip_probs = std::move(skip_probs); + fc.joints = std::move(joints); + fc.sign = std::move(sign); + fc.class_0 = std::move(class_0); + fc.fr = std::move(fr); + fc.class_0_hp = std::move(class_0_hp); + fc.high_precision = std::move(high_precision); + fc.classes = std::move(classes); + fc.class_0_fr = std::move(class_0_fr); + fc.prob_bits = std::move(pred_bits); + fc.single_ref_prob = std::move(single_ref_prob); + fc.comp_ref_prob = std::move(comp_ref_prob); + + // Skip the 4th element as it goes unused + for (std::size_t i = 0; i < coef_probs.size(); i += 4) { + const std::size_t j = i - i / 4; + fc.coef_probs[j] = coef_probs[i]; + fc.coef_probs[j + 1] = coef_probs[i + 1]; + fc.coef_probs[j + 2] = coef_probs[i + 2]; + } } }; static_assert(sizeof(EntropyProbs) == 0xEA0, "EntropyProbs is an invalid size"); diff --git a/src/video_core/command_classes/nvdec.cpp b/src/video_core/command_classes/nvdec.cpp index 8ca7a7b06f..79e1f4e137 100644 --- a/src/video_core/command_classes/nvdec.cpp +++ b/src/video_core/command_classes/nvdec.cpp @@ -29,11 +29,7 @@ void Nvdec::ProcessMethod(Method method, const std::vector& arguments) { } } -AVFrame* Nvdec::GetFrame() { - return codec->GetCurrentFrame(); -} - -const AVFrame* Nvdec::GetFrame() const { +AVFramePtr Nvdec::GetFrame() { return codec->GetCurrentFrame(); } diff --git a/src/video_core/command_classes/nvdec.h b/src/video_core/command_classes/nvdec.h index eec4443f9f..e4877c5330 100644 --- a/src/video_core/command_classes/nvdec.h +++ b/src/video_core/command_classes/nvdec.h @@ -26,8 +26,7 @@ public: void ProcessMethod(Method method, const std::vector& arguments); /// Return most recently decoded frame - [[nodiscard]] AVFrame* GetFrame(); - [[nodiscard]] const AVFrame* GetFrame() const; + [[nodiscard]] AVFramePtr GetFrame(); private: /// Invoke codec to decode a frame diff --git a/src/video_core/command_classes/vic.cpp b/src/video_core/command_classes/vic.cpp index 5b52da277f..2484430279 100644 --- a/src/video_core/command_classes/vic.cpp +++ b/src/video_core/command_classes/vic.cpp @@ -58,17 +58,18 @@ void Vic::Execute() { return; } const VicConfig config{gpu.MemoryManager().Read(config_struct_address + 0x20)}; + const AVFramePtr frame_ptr = std::move(nvdec_processor->GetFrame()); + const auto* frame = frame_ptr.get(); + if (!frame || frame->width == 0 || frame->height == 0) { + return; + } const VideoPixelFormat pixel_format = static_cast(config.pixel_format.Value()); switch (pixel_format) { case VideoPixelFormat::BGRA8: case VideoPixelFormat::RGBA8: { LOG_TRACE(Service_NVDRV, "Writing RGB Frame"); - const auto* frame = nvdec_processor->GetFrame(); - if (!frame || frame->width == 0 || frame->height == 0) { - return; - } if (scaler_ctx == nullptr || frame->width != scaler_width || frame->height != scaler_height) { const AVPixelFormat target_format = @@ -121,12 +122,6 @@ void Vic::Execute() { case VideoPixelFormat::Yuv420: { LOG_TRACE(Service_NVDRV, "Writing YUV420 Frame"); - const auto* frame = nvdec_processor->GetFrame(); - - if (!frame || frame->width == 0 || frame->height == 0) { - return; - } - const std::size_t surface_width = config.surface_width_minus1 + 1; const std::size_t surface_height = config.surface_height_minus1 + 1; const std::size_t half_width = surface_width / 2; -- cgit v1.2.3 From aa2d90b35cc900f39c1cd1ae516a2dc87364429d Mon Sep 17 00:00:00 2001 From: ameerj Date: Thu, 26 Nov 2020 00:18:26 -0500 Subject: Address PR feedback remove some redundant moves, make deleter match naming guidelines. Co-Authored-By: LC <712067+lioncash@users.noreply.github.com> --- src/video_core/command_classes/codecs/codec.cpp | 15 ++++---- src/video_core/command_classes/codecs/codec.h | 4 +-- src/video_core/command_classes/codecs/vp9.cpp | 2 +- src/video_core/command_classes/codecs/vp9_types.h | 44 +++++++++++------------ 4 files changed, 33 insertions(+), 32 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/codec.cpp b/src/video_core/command_classes/codecs/codec.cpp index 1a19341c85..412e1e41cc 100644 --- a/src/video_core/command_classes/codecs/codec.cpp +++ b/src/video_core/command_classes/codecs/codec.cpp @@ -18,7 +18,7 @@ extern "C" { namespace Tegra { -void av_frame_deleter(AVFrame* ptr) { +void AVFrameDeleter(AVFrame* ptr) { av_frame_unref(ptr); av_free(ptr); } @@ -101,7 +101,7 @@ void Codec::Decode() { if (!vp9_hidden_frame) { // Only receive/store visible frames - AVFramePtr frame = AVFramePtr{av_frame_alloc(), av_frame_deleter}; + AVFramePtr frame = AVFramePtr{av_frame_alloc(), AVFrameDeleter}; avcodec_receive_frame(av_codec_ctx, frame.get()); av_frames.push(std::move(frame)); } @@ -110,12 +110,13 @@ void Codec::Decode() { AVFramePtr Codec::GetCurrentFrame() { // Sometimes VIC will request more frames than have been decoded. // in this case, return a nullptr and don't overwrite previous frame data - if (av_frames.size() > 0) { - AVFramePtr frame = std::move(av_frames.front()); - av_frames.pop(); - return frame; + if (av_frames.empty()) { + return AVFramePtr{nullptr, AVFrameDeleter}; } - return AVFramePtr{nullptr, av_frame_deleter}; + + AVFramePtr frame = std::move(av_frames.front()); + av_frames.pop(); + return frame; } NvdecCommon::VideoCodec Codec::GetCurrentCodec() const { diff --git a/src/video_core/command_classes/codecs/codec.h b/src/video_core/command_classes/codecs/codec.h index c26b59fdec..0c6dde4058 100644 --- a/src/video_core/command_classes/codecs/codec.h +++ b/src/video_core/command_classes/codecs/codec.h @@ -23,8 +23,8 @@ namespace Tegra { class GPU; struct VicRegisters; -void av_frame_deleter(AVFrame* ptr); -using AVFramePtr = std::unique_ptr; +void AVFrameDeleter(AVFrame* ptr); +using AVFramePtr = std::unique_ptr; namespace Decoder { class H264; diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 31e00c27d7..b1d675cdbc 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -306,7 +306,7 @@ void VP9::WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, const std::array& old_prob) { constexpr u32 block_bytes = 2 * 2 * 6 * 6 * 3; - const auto needs_update = [&](u32 base_index) -> bool { + const auto needs_update = [&](u32 base_index) { return !std::equal(new_prob.begin() + base_index, new_prob.begin() + base_index + block_bytes, old_prob.begin() + base_index); diff --git a/src/video_core/command_classes/codecs/vp9_types.h b/src/video_core/command_classes/codecs/vp9_types.h index 5ca944f2ac..139501a1c6 100644 --- a/src/video_core/command_classes/codecs/vp9_types.h +++ b/src/video_core/command_classes/codecs/vp9_types.h @@ -245,33 +245,33 @@ struct EntropyProbs { std::array coef_probs{}; void Convert(Vp9EntropyProbs& fc) { - fc.inter_mode_prob = std::move(inter_mode_prob); - fc.intra_inter_prob = std::move(intra_inter_prob); - fc.tx_8x8_prob = std::move(tx_8x8_prob); - fc.tx_16x16_prob = std::move(tx_16x16_prob); - fc.tx_32x32_prob = std::move(tx_32x32_prob); + fc.inter_mode_prob = inter_mode_prob; + fc.intra_inter_prob = intra_inter_prob; + fc.tx_8x8_prob = tx_8x8_prob; + fc.tx_16x16_prob = tx_16x16_prob; + fc.tx_32x32_prob = tx_32x32_prob; - for (s32 i = 0; i < 4; i++) { - for (s32 j = 0; j < 9; j++) { + for (std::size_t i = 0; i < 4; i++) { + for (std::size_t j = 0; j < 9; j++) { fc.y_mode_prob[j + 9 * i] = j < 8 ? y_mode_prob_e0e7[i][j] : y_mode_prob_e8[i]; } } - fc.partition_prob = std::move(partition_prob); - fc.switchable_interp_prob = std::move(switchable_interp_prob); - fc.comp_inter_prob = std::move(comp_inter_prob); - fc.skip_probs = std::move(skip_probs); - fc.joints = std::move(joints); - fc.sign = std::move(sign); - fc.class_0 = std::move(class_0); - fc.fr = std::move(fr); - fc.class_0_hp = std::move(class_0_hp); - fc.high_precision = std::move(high_precision); - fc.classes = std::move(classes); - fc.class_0_fr = std::move(class_0_fr); - fc.prob_bits = std::move(pred_bits); - fc.single_ref_prob = std::move(single_ref_prob); - fc.comp_ref_prob = std::move(comp_ref_prob); + fc.partition_prob = partition_prob; + fc.switchable_interp_prob = switchable_interp_prob; + fc.comp_inter_prob = comp_inter_prob; + fc.skip_probs = skip_probs; + fc.joints = joints; + fc.sign = sign; + fc.class_0 = class_0; + fc.fr = fr; + fc.class_0_hp = class_0_hp; + fc.high_precision = high_precision; + fc.classes = classes; + fc.class_0_fr = class_0_fr; + fc.prob_bits = pred_bits; + fc.single_ref_prob = single_ref_prob; + fc.comp_ref_prob = comp_ref_prob; // Skip the 4th element as it goes unused for (std::size_t i = 0; i < coef_probs.size(); i += 4) { -- cgit v1.2.3 From 92a5b7b388fcb837d9701ce9639c9fc4fc78422b Mon Sep 17 00:00:00 2001 From: Lioncash Date: Thu, 3 Dec 2020 12:33:05 -0500 Subject: vp9/vic: Resolve pessimizing moves Removes the usage of moves that don't result in behavior different from a copy, or otherwise would prevent copy elision from occurring. --- src/video_core/command_classes/codecs/vp9.cpp | 20 ++++++++++---------- src/video_core/command_classes/vic.cpp | 2 +- 2 files changed, 11 insertions(+), 11 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index b1d675cdbc..7d8d6ee3ca 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -355,7 +355,7 @@ void VP9::WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_ Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) { PictureInfo picture_info{}; gpu.MemoryManager().ReadBlock(state.picture_info_offset, &picture_info, sizeof(PictureInfo)); - Vp9PictureInfo vp9_info = std::move(picture_info.Convert()); + Vp9PictureInfo vp9_info = picture_info.Convert(); InsertEntropy(state.vp9_entropy_probs_offset, vp9_info.entropy); @@ -377,7 +377,7 @@ Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) Vp9FrameContainer frame{}; { gpu.SyncGuestHost(); - frame.info = std::move(GetVp9PictureInfo(state)); + frame.info = GetVp9PictureInfo(state); frame.bit_stream.resize(frame.info.bitstream_size); gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, frame.bit_stream.data(), frame.info.bitstream_size); @@ -385,29 +385,29 @@ Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) // Buffer two frames, saving the last show frame info if (!next_next_frame.bit_stream.empty()) { Vp9FrameContainer temp{ - .info = std::move(frame.info), + .info = frame.info, .bit_stream = std::move(frame.bit_stream), }; next_next_frame.info.show_frame = frame.info.last_frame_shown; - frame.info = std::move(next_next_frame.info); + frame.info = next_next_frame.info; frame.bit_stream = std::move(next_next_frame.bit_stream); next_next_frame = std::move(temp); if (!next_frame.bit_stream.empty()) { Vp9FrameContainer temp2{ - .info = std::move(frame.info), + .info = frame.info, .bit_stream = std::move(frame.bit_stream), }; next_frame.info.show_frame = frame.info.last_frame_shown; - frame.info = std::move(next_frame.info); + frame.info = next_frame.info; frame.bit_stream = std::move(next_frame.bit_stream); next_frame = std::move(temp2); } else { - next_frame.info = std::move(frame.info); + next_frame.info = frame.info; next_frame.bit_stream = std::move(frame.bit_stream); } } else { - next_next_frame.info = std::move(frame.info); + next_next_frame.info = frame.info; next_next_frame.bit_stream = std::move(frame.bit_stream); } return frame; @@ -806,8 +806,8 @@ VpxBitStreamWriter VP9::ComposeUncompressedHeader() { const std::vector& VP9::ComposeFrameHeader(const NvdecCommon::NvdecRegisters& state) { std::vector bitstream; { - Vp9FrameContainer curr_frame = std::move(GetCurrentFrame(state)); - current_frame_info = std::move(curr_frame.info); + Vp9FrameContainer curr_frame = GetCurrentFrame(state); + current_frame_info = curr_frame.info; bitstream = std::move(curr_frame.bit_stream); } diff --git a/src/video_core/command_classes/vic.cpp b/src/video_core/command_classes/vic.cpp index 2484430279..6cfc193fa9 100644 --- a/src/video_core/command_classes/vic.cpp +++ b/src/video_core/command_classes/vic.cpp @@ -58,7 +58,7 @@ void Vic::Execute() { return; } const VicConfig config{gpu.MemoryManager().Read(config_struct_address + 0x20)}; - const AVFramePtr frame_ptr = std::move(nvdec_processor->GetFrame()); + const AVFramePtr frame_ptr = nvdec_processor->GetFrame(); const auto* frame = frame_ptr.get(); if (!frame || frame->width == 0 || frame->height == 0) { return; -- cgit v1.2.3 From 2f181b6a90cb3be8b4eed32601c619dc385d18e8 Mon Sep 17 00:00:00 2001 From: Lioncash Date: Fri, 4 Dec 2020 14:39:12 -0500 Subject: video_core: Resolve more variable shadowing scenarios 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. --- src/video_core/cdma_pusher.cpp | 30 +++++++------- src/video_core/cdma_pusher.h | 16 ++++---- src/video_core/command_classes/codecs/vp9.cpp | 2 +- src/video_core/command_classes/codecs/vp9.h | 2 +- src/video_core/dma_pusher.cpp | 9 +++- src/video_core/dma_pusher.h | 14 +++---- src/video_core/engines/engine_upload.cpp | 8 ++-- src/video_core/engines/engine_upload.h | 4 +- src/video_core/engines/kepler_memory.cpp | 4 +- src/video_core/engines/kepler_memory.h | 2 +- src/video_core/engines/maxwell_dma.cpp | 6 ++- src/video_core/engines/maxwell_dma.h | 4 +- src/video_core/fence_manager.h | 8 ++-- src/video_core/gpu.cpp | 8 +++- src/video_core/gpu.h | 12 +++--- src/video_core/gpu_asynch.cpp | 4 +- src/video_core/gpu_asynch.h | 2 +- src/video_core/gpu_synch.cpp | 2 +- src/video_core/gpu_synch.h | 2 +- src/video_core/gpu_thread.cpp | 16 ++++---- src/video_core/gpu_thread.h | 25 ++++++----- src/video_core/guest_driver.h | 4 +- src/video_core/macro/macro_hle.cpp | 6 +-- src/video_core/macro/macro_hle.h | 2 +- src/video_core/macro/macro_interpreter.cpp | 20 ++++----- src/video_core/macro/macro_interpreter.h | 10 ++--- src/video_core/macro/macro_jit_x64.cpp | 16 ++++---- src/video_core/macro/macro_jit_x64.h | 4 +- src/video_core/memory_manager.h | 4 +- .../renderer_opengl/gl_arb_decompiler.cpp | 48 +++++++++++----------- .../renderer_opengl/gl_fence_manager.cpp | 14 +++---- src/video_core/renderer_opengl/gl_fence_manager.h | 10 ++--- src/video_core/renderer_opengl/gl_query_cache.cpp | 15 +++---- src/video_core/renderer_opengl/gl_query_cache.h | 6 +-- src/video_core/renderer_opengl/gl_shader_cache.cpp | 6 +-- src/video_core/renderer_opengl/gl_shader_cache.h | 2 +- .../renderer_opengl/gl_shader_decompiler.cpp | 35 ++++++++-------- .../renderer_opengl/gl_shader_decompiler.h | 12 +++--- .../renderer_opengl/gl_texture_cache.cpp | 20 ++++----- src/video_core/renderer_opengl/gl_texture_cache.h | 5 ++- src/video_core/renderer_vulkan/vk_command_pool.cpp | 4 +- src/video_core/renderer_vulkan/vk_command_pool.h | 2 +- 42 files changed, 219 insertions(+), 206 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/cdma_pusher.cpp b/src/video_core/cdma_pusher.cpp index b60f862603..e3e7432f74 100644 --- a/src/video_core/cdma_pusher.cpp +++ b/src/video_core/cdma_pusher.cpp @@ -29,8 +29,8 @@ #include "video_core/memory_manager.h" namespace Tegra { -CDmaPusher::CDmaPusher(GPU& gpu) - : gpu(gpu), nvdec_processor(std::make_shared(gpu)), +CDmaPusher::CDmaPusher(GPU& gpu_) + : gpu{gpu_}, nvdec_processor(std::make_shared(gpu)), vic_processor(std::make_unique(gpu, nvdec_processor)), host1x_processor(std::make_unique(gpu)), nvdec_sync(std::make_unique(gpu)), @@ -100,11 +100,11 @@ void CDmaPusher::Step() { } } -void CDmaPusher::ExecuteCommand(u32 offset, u32 data) { +void CDmaPusher::ExecuteCommand(u32 state_offset, u32 data) { switch (current_class) { case ChClassId::NvDec: - ThiStateWrite(nvdec_thi_state, offset, {data}); - switch (static_cast(offset)) { + ThiStateWrite(nvdec_thi_state, state_offset, {data}); + switch (static_cast(state_offset)) { case ThiMethod::IncSyncpt: { LOG_DEBUG(Service_NVDRV, "NVDEC Class IncSyncpt Method"); const auto syncpoint_id = static_cast(data & 0xFF); @@ -120,16 +120,16 @@ void CDmaPusher::ExecuteCommand(u32 offset, u32 data) { case ThiMethod::SetMethod1: LOG_DEBUG(Service_NVDRV, "NVDEC method 0x{:X}", static_cast(nvdec_thi_state.method_0)); - nvdec_processor->ProcessMethod( - static_cast(nvdec_thi_state.method_0), {data}); + nvdec_processor->ProcessMethod(static_cast(nvdec_thi_state.method_0), + {data}); break; default: break; } break; case ChClassId::GraphicsVic: - ThiStateWrite(vic_thi_state, static_cast(offset), {data}); - switch (static_cast(offset)) { + ThiStateWrite(vic_thi_state, static_cast(state_offset), {data}); + switch (static_cast(state_offset)) { case ThiMethod::IncSyncpt: { LOG_DEBUG(Service_NVDRV, "VIC Class IncSyncpt Method"); const auto syncpoint_id = static_cast(data & 0xFF); @@ -145,8 +145,7 @@ void CDmaPusher::ExecuteCommand(u32 offset, u32 data) { case ThiMethod::SetMethod1: LOG_DEBUG(Service_NVDRV, "VIC method 0x{:X}, Args=({})", static_cast(vic_thi_state.method_0), data); - vic_processor->ProcessMethod(static_cast(vic_thi_state.method_0), - {data}); + vic_processor->ProcessMethod(static_cast(vic_thi_state.method_0), {data}); break; default: break; @@ -155,7 +154,7 @@ void CDmaPusher::ExecuteCommand(u32 offset, u32 data) { case ChClassId::Host1x: // This device is mainly for syncpoint synchronization LOG_DEBUG(Service_NVDRV, "Host1X Class Method"); - host1x_processor->ProcessMethod(static_cast(offset), {data}); + host1x_processor->ProcessMethod(static_cast(state_offset), {data}); break; default: UNIMPLEMENTED_MSG("Current class not implemented {:X}", static_cast(current_class)); @@ -163,9 +162,10 @@ void CDmaPusher::ExecuteCommand(u32 offset, u32 data) { } } -void CDmaPusher::ThiStateWrite(ThiRegisters& state, u32 offset, const std::vector& arguments) { - u8* const state_offset = reinterpret_cast(&state) + sizeof(u32) * offset; - std::memcpy(state_offset, arguments.data(), sizeof(u32) * arguments.size()); +void CDmaPusher::ThiStateWrite(ThiRegisters& state, u32 state_offset, + const std::vector& arguments) { + u8* const state_offset_ptr = reinterpret_cast(&state) + sizeof(u32) * state_offset; + std::memcpy(state_offset_ptr, arguments.data(), sizeof(u32) * arguments.size()); } } // namespace Tegra diff --git a/src/video_core/cdma_pusher.h b/src/video_core/cdma_pusher.h index 982f309c53..0db1cd6467 100644 --- a/src/video_core/cdma_pusher.h +++ b/src/video_core/cdma_pusher.h @@ -68,8 +68,8 @@ struct ChCommand { std::vector arguments; }; -using ChCommandHeaderList = std::vector; -using ChCommandList = std::vector; +using ChCommandHeaderList = std::vector; +using ChCommandList = std::vector; struct ThiRegisters { u32_le increment_syncpt{}; @@ -96,7 +96,7 @@ enum class ThiMethod : u32 { class CDmaPusher { public: - explicit CDmaPusher(GPU& gpu); + explicit CDmaPusher(GPU& gpu_); ~CDmaPusher(); /// Push NVDEC command buffer entries into queue @@ -109,17 +109,17 @@ public: void Step(); /// Invoke command class devices to execute the command based on the current state - void ExecuteCommand(u32 offset, u32 data); + void ExecuteCommand(u32 state_offset, u32 data); private: /// Write arguments value to the ThiRegisters member at the specified offset - void ThiStateWrite(ThiRegisters& state, u32 offset, const std::vector& arguments); + void ThiStateWrite(ThiRegisters& state, u32 state_offset, const std::vector& arguments); GPU& gpu; - std::shared_ptr nvdec_processor; - std::unique_ptr vic_processor; - std::unique_ptr host1x_processor; + std::shared_ptr nvdec_processor; + std::unique_ptr vic_processor; + std::unique_ptr host1x_processor; std::unique_ptr nvdec_sync; std::unique_ptr vic_sync; ChClassId current_class{}; diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 7d8d6ee3ca..1771bc939b 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -233,7 +233,7 @@ constexpr std::array map_lut{ } } // Anonymous namespace -VP9::VP9(GPU& gpu) : gpu(gpu) {} +VP9::VP9(GPU& gpu_) : gpu{gpu_} {} VP9::~VP9() = default; diff --git a/src/video_core/command_classes/codecs/vp9.h b/src/video_core/command_classes/codecs/vp9.h index 9ebbbf59e5..8396c81054 100644 --- a/src/video_core/command_classes/codecs/vp9.h +++ b/src/video_core/command_classes/codecs/vp9.h @@ -108,7 +108,7 @@ private: class VP9 { public: - explicit VP9(GPU& gpu); + explicit VP9(GPU& gpu_); ~VP9(); VP9(const VP9&) = delete; diff --git a/src/video_core/dma_pusher.cpp b/src/video_core/dma_pusher.cpp index d8801b1f5a..2c8b200241 100644 --- a/src/video_core/dma_pusher.cpp +++ b/src/video_core/dma_pusher.cpp @@ -13,7 +13,7 @@ namespace Tegra { -DmaPusher::DmaPusher(Core::System& system, GPU& gpu) : gpu{gpu}, system{system} {} +DmaPusher::DmaPusher(Core::System& system_, GPU& gpu_) : gpu{gpu_}, system{system_} {} DmaPusher::~DmaPusher() = default; @@ -152,7 +152,12 @@ void DmaPusher::SetState(const CommandHeader& command_header) { void DmaPusher::CallMethod(u32 argument) const { if (dma_state.method < non_puller_methods) { - gpu.CallMethod({dma_state.method, argument, dma_state.subchannel, dma_state.method_count}); + gpu.CallMethod(GPU::MethodCall{ + dma_state.method, + argument, + dma_state.subchannel, + dma_state.method_count, + }); } else { subchannels[dma_state.subchannel]->CallMethod(dma_state.method, argument, dma_state.is_last_call); diff --git a/src/video_core/dma_pusher.h b/src/video_core/dma_pusher.h index 96ac267f75..19f286fa71 100644 --- a/src/video_core/dma_pusher.h +++ b/src/video_core/dma_pusher.h @@ -87,11 +87,11 @@ inline CommandHeader BuildCommandHeader(BufferMethods method, u32 arg_count, Sub struct CommandList final { CommandList() = default; explicit CommandList(std::size_t size) : command_lists(size) {} - explicit CommandList(std::vector&& prefetch_command_list) - : prefetch_command_list{std::move(prefetch_command_list)} {} + explicit CommandList(std::vector&& prefetch_command_list_) + : prefetch_command_list{std::move(prefetch_command_list_)} {} - std::vector command_lists; - std::vector prefetch_command_list; + std::vector command_lists; + std::vector prefetch_command_list; }; /** @@ -103,7 +103,7 @@ struct CommandList final { */ class DmaPusher final { public: - explicit DmaPusher(Core::System& system, GPU& gpu); + explicit DmaPusher(Core::System& system_, GPU& gpu_); ~DmaPusher(); void Push(CommandList&& entries) { @@ -112,7 +112,7 @@ public: void DispatchCalls(); - void BindSubchannel(Tegra::Engines::EngineInterface* engine, u32 subchannel_id) { + void BindSubchannel(Engines::EngineInterface* engine, u32 subchannel_id) { subchannels[subchannel_id] = engine; } @@ -145,7 +145,7 @@ private: bool ib_enable{true}; ///< IB mode enabled - std::array subchannels{}; + std::array subchannels{}; GPU& gpu; Core::System& system; diff --git a/src/video_core/engines/engine_upload.cpp b/src/video_core/engines/engine_upload.cpp index d44ad0cd8d..71d7e1473f 100644 --- a/src/video_core/engines/engine_upload.cpp +++ b/src/video_core/engines/engine_upload.cpp @@ -11,16 +11,16 @@ namespace Tegra::Engines::Upload { -State::State(MemoryManager& memory_manager, Registers& regs) - : regs{regs}, memory_manager{memory_manager} {} +State::State(MemoryManager& memory_manager_, Registers& regs_) + : regs{regs_}, memory_manager{memory_manager_} {} State::~State() = default; -void State::ProcessExec(const bool is_linear) { +void State::ProcessExec(const bool is_linear_) { write_offset = 0; copy_size = regs.line_length_in * regs.line_count; inner_buffer.resize(copy_size); - this->is_linear = is_linear; + is_linear = is_linear_; } void State::ProcessData(const u32 data, const bool is_last_call) { diff --git a/src/video_core/engines/engine_upload.h b/src/video_core/engines/engine_upload.h index 462da419e2..1c7f1effac 100644 --- a/src/video_core/engines/engine_upload.h +++ b/src/video_core/engines/engine_upload.h @@ -54,10 +54,10 @@ struct Registers { class State { public: - State(MemoryManager& memory_manager, Registers& regs); + explicit State(MemoryManager& memory_manager_, Registers& regs_); ~State(); - void ProcessExec(bool is_linear); + void ProcessExec(bool is_linear_); void ProcessData(u32 data, bool is_last_call); private: diff --git a/src/video_core/engines/kepler_memory.cpp b/src/video_core/engines/kepler_memory.cpp index dc71b2eec3..9911140e93 100644 --- a/src/video_core/engines/kepler_memory.cpp +++ b/src/video_core/engines/kepler_memory.cpp @@ -14,8 +14,8 @@ namespace Tegra::Engines { -KeplerMemory::KeplerMemory(Core::System& system, MemoryManager& memory_manager) - : system{system}, upload_state{memory_manager, regs.upload} {} +KeplerMemory::KeplerMemory(Core::System& system_, MemoryManager& memory_manager) + : system{system_}, upload_state{memory_manager, regs.upload} {} KeplerMemory::~KeplerMemory() = default; diff --git a/src/video_core/engines/kepler_memory.h b/src/video_core/engines/kepler_memory.h index 5b7f71a00b..62483589e6 100644 --- a/src/video_core/engines/kepler_memory.h +++ b/src/video_core/engines/kepler_memory.h @@ -35,7 +35,7 @@ namespace Tegra::Engines { class KeplerMemory final : public EngineInterface { public: - KeplerMemory(Core::System& system, MemoryManager& memory_manager); + explicit KeplerMemory(Core::System& system_, MemoryManager& memory_manager); ~KeplerMemory(); /// Write the value to the register identified by method. diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp index 8fa359d0a6..1c29e895ee 100644 --- a/src/video_core/engines/maxwell_dma.cpp +++ b/src/video_core/engines/maxwell_dma.cpp @@ -16,8 +16,10 @@ namespace Tegra::Engines { using namespace Texture; -MaxwellDMA::MaxwellDMA(Core::System& system, MemoryManager& memory_manager) - : system{system}, memory_manager{memory_manager} {} +MaxwellDMA::MaxwellDMA(Core::System& system_, MemoryManager& memory_manager_) + : system{system_}, memory_manager{memory_manager_} {} + +MaxwellDMA::~MaxwellDMA() = default; void MaxwellDMA::CallMethod(u32 method, u32 method_argument, bool is_last_call) { ASSERT_MSG(method < NUM_REGS, "Invalid MaxwellDMA register"); diff --git a/src/video_core/engines/maxwell_dma.h b/src/video_core/engines/maxwell_dma.h index 50f445efcd..17bd280c4f 100644 --- a/src/video_core/engines/maxwell_dma.h +++ b/src/video_core/engines/maxwell_dma.h @@ -185,8 +185,8 @@ public: }; static_assert(sizeof(RemapConst) == 12); - explicit MaxwellDMA(Core::System& system, MemoryManager& memory_manager); - ~MaxwellDMA() = default; + explicit MaxwellDMA(Core::System& system_, MemoryManager& memory_manager_); + ~MaxwellDMA(); /// Write the value to the register identified by method. void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; diff --git a/src/video_core/fence_manager.h b/src/video_core/fence_manager.h index de6991ef65..c5f26896e4 100644 --- a/src/video_core/fence_manager.h +++ b/src/video_core/fence_manager.h @@ -17,11 +17,11 @@ namespace VideoCommon { class FenceBase { public: - FenceBase(u32 payload, bool is_stubbed) - : address{}, payload{payload}, is_semaphore{false}, is_stubbed{is_stubbed} {} + explicit FenceBase(u32 payload_, bool is_stubbed_) + : address{}, payload{payload_}, is_semaphore{false}, is_stubbed{is_stubbed_} {} - FenceBase(GPUVAddr address, u32 payload, bool is_stubbed) - : address{address}, payload{payload}, is_semaphore{true}, is_stubbed{is_stubbed} {} + explicit FenceBase(GPUVAddr address_, u32 payload_, bool is_stubbed_) + : address{address_}, payload{payload_}, is_semaphore{true}, is_stubbed{is_stubbed_} {} GPUVAddr GetAddress() const { return address; diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp index e91f52938c..964b3f3dcd 100644 --- a/src/video_core/gpu.cpp +++ b/src/video_core/gpu.cpp @@ -232,8 +232,12 @@ void GPU::CallMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 CallEngineMultiMethod(method, subchannel, base_start, amount, methods_pending); } else { for (std::size_t i = 0; i < amount; i++) { - CallPullerMethod( - {method, base_start[i], subchannel, methods_pending - static_cast(i)}); + CallPullerMethod(MethodCall{ + method, + base_start[i], + subchannel, + methods_pending - static_cast(i), + }); } } } diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h index 21410e125d..660641d041 100644 --- a/src/video_core/gpu.h +++ b/src/video_core/gpu.h @@ -149,16 +149,16 @@ public: u32 subchannel{}; u32 method_count{}; - MethodCall(u32 method, u32 argument, u32 subchannel = 0, u32 method_count = 0) - : method(method), argument(argument), subchannel(subchannel), - method_count(method_count) {} + explicit MethodCall(u32 method_, u32 argument_, u32 subchannel_ = 0, u32 method_count_ = 0) + : method(method_), argument(argument_), subchannel(subchannel_), + method_count(method_count_) {} [[nodiscard]] bool IsLastCall() const { return method_count <= 1; } }; - explicit GPU(Core::System& system, bool is_async, bool use_nvdec); + explicit GPU(Core::System& system_, bool is_async_, bool use_nvdec_); virtual ~GPU(); /// Binds a renderer to the GPU. @@ -414,8 +414,8 @@ private: std::condition_variable sync_cv; struct FlushRequest { - FlushRequest(u64 fence, VAddr addr, std::size_t size) - : fence{fence}, addr{addr}, size{size} {} + explicit FlushRequest(u64 fence_, VAddr addr_, std::size_t size_) + : fence{fence_}, addr{addr_}, size{size_} {} u64 fence; VAddr addr; std::size_t size; diff --git a/src/video_core/gpu_asynch.cpp b/src/video_core/gpu_asynch.cpp index a9baaf7ef0..6cc091ecd2 100644 --- a/src/video_core/gpu_asynch.cpp +++ b/src/video_core/gpu_asynch.cpp @@ -10,8 +10,8 @@ namespace VideoCommon { -GPUAsynch::GPUAsynch(Core::System& system, bool use_nvdec) - : GPU{system, true, use_nvdec}, gpu_thread{system} {} +GPUAsynch::GPUAsynch(Core::System& system_, bool use_nvdec_) + : GPU{system_, true, use_nvdec_}, gpu_thread{system_} {} GPUAsynch::~GPUAsynch() = default; diff --git a/src/video_core/gpu_asynch.h b/src/video_core/gpu_asynch.h index 0c0872e73f..a384113f4b 100644 --- a/src/video_core/gpu_asynch.h +++ b/src/video_core/gpu_asynch.h @@ -20,7 +20,7 @@ namespace VideoCommon { /// Implementation of GPU interface that runs the GPU asynchronously class GPUAsynch final : public Tegra::GPU { public: - explicit GPUAsynch(Core::System& system, bool use_nvdec); + explicit GPUAsynch(Core::System& system_, bool use_nvdec_); ~GPUAsynch() override; void Start() override; diff --git a/src/video_core/gpu_synch.cpp b/src/video_core/gpu_synch.cpp index ecf7bbdf3b..1e9d4b9b2c 100644 --- a/src/video_core/gpu_synch.cpp +++ b/src/video_core/gpu_synch.cpp @@ -7,7 +7,7 @@ namespace VideoCommon { -GPUSynch::GPUSynch(Core::System& system, bool use_nvdec) : GPU{system, false, use_nvdec} {} +GPUSynch::GPUSynch(Core::System& system_, bool use_nvdec_) : GPU{system_, false, use_nvdec_} {} GPUSynch::~GPUSynch() = default; diff --git a/src/video_core/gpu_synch.h b/src/video_core/gpu_synch.h index 9d778c71a0..c5904b8db8 100644 --- a/src/video_core/gpu_synch.h +++ b/src/video_core/gpu_synch.h @@ -19,7 +19,7 @@ namespace VideoCommon { /// Implementation of GPU interface that runs the GPU synchronously class GPUSynch final : public Tegra::GPU { public: - explicit GPUSynch(Core::System& system, bool use_nvdec); + explicit GPUSynch(Core::System& system_, bool use_nvdec_); ~GPUSynch() override; void Start() override; diff --git a/src/video_core/gpu_thread.cpp b/src/video_core/gpu_thread.cpp index 4b8f582835..e27218b963 100644 --- a/src/video_core/gpu_thread.cpp +++ b/src/video_core/gpu_thread.cpp @@ -39,23 +39,23 @@ static void RunThread(Core::System& system, VideoCore::RendererBase& renderer, CommandDataContainer next; while (state.is_running) { next = state.queue.PopWait(); - if (const auto submit_list = std::get_if(&next.data)) { + if (auto* submit_list = std::get_if(&next.data)) { dma_pusher.Push(std::move(submit_list->entries)); dma_pusher.DispatchCalls(); - } else if (const auto command_list = std::get_if(&next.data)) { + } else if (auto* command_list = std::get_if(&next.data)) { // NVDEC cdma_pusher.Push(std::move(command_list->entries)); cdma_pusher.DispatchCalls(); - } else if (const auto data = std::get_if(&next.data)) { + } else if (const auto* data = std::get_if(&next.data)) { renderer.SwapBuffers(data->framebuffer ? &*data->framebuffer : nullptr); } else if (std::holds_alternative(next.data)) { renderer.Rasterizer().ReleaseFences(); } else if (std::holds_alternative(next.data)) { system.GPU().TickWork(); - } else if (const auto data = std::get_if(&next.data)) { - renderer.Rasterizer().FlushRegion(data->addr, data->size); - } else if (const auto data = std::get_if(&next.data)) { - renderer.Rasterizer().OnCPUWrite(data->addr, data->size); + } else if (const auto* flush = std::get_if(&next.data)) { + renderer.Rasterizer().FlushRegion(flush->addr, flush->size); + } else if (const auto* invalidate = std::get_if(&next.data)) { + renderer.Rasterizer().OnCPUWrite(invalidate->addr, invalidate->size); } else if (std::holds_alternative(next.data)) { return; } else { @@ -65,7 +65,7 @@ static void RunThread(Core::System& system, VideoCore::RendererBase& renderer, } } -ThreadManager::ThreadManager(Core::System& system) : system{system} {} +ThreadManager::ThreadManager(Core::System& system_) : system{system_} {} ThreadManager::~ThreadManager() { if (!thread.joinable()) { diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index 32a34e3a72..f1c52cd9e0 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -32,30 +32,30 @@ struct EndProcessingCommand final {}; /// Command to signal to the GPU thread that a command list is ready for processing struct SubmitListCommand final { - explicit SubmitListCommand(Tegra::CommandList&& entries) : entries{std::move(entries)} {} + explicit SubmitListCommand(Tegra::CommandList&& entries_) : entries{std::move(entries_)} {} Tegra::CommandList entries; }; /// Command to signal to the GPU thread that a cdma command list is ready for processing struct SubmitChCommandEntries final { - explicit SubmitChCommandEntries(Tegra::ChCommandHeaderList&& entries) - : entries{std::move(entries)} {} + explicit SubmitChCommandEntries(Tegra::ChCommandHeaderList&& entries_) + : entries{std::move(entries_)} {} Tegra::ChCommandHeaderList entries; }; /// Command to signal to the GPU thread that a swap buffers is pending struct SwapBuffersCommand final { - explicit SwapBuffersCommand(std::optional framebuffer) - : framebuffer{std::move(framebuffer)} {} + explicit SwapBuffersCommand(std::optional framebuffer_) + : framebuffer{std::move(framebuffer_)} {} std::optional framebuffer; }; /// Command to signal to the GPU thread to flush a region struct FlushRegionCommand final { - explicit constexpr FlushRegionCommand(VAddr addr, u64 size) : addr{addr}, size{size} {} + explicit constexpr FlushRegionCommand(VAddr addr_, u64 size_) : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -63,7 +63,7 @@ struct FlushRegionCommand final { /// Command to signal to the GPU thread to invalidate a region struct InvalidateRegionCommand final { - explicit constexpr InvalidateRegionCommand(VAddr addr, u64 size) : addr{addr}, size{size} {} + explicit constexpr InvalidateRegionCommand(VAddr addr_, u64 size_) : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -71,8 +71,8 @@ struct InvalidateRegionCommand final { /// Command to signal to the GPU thread to flush and invalidate a region struct FlushAndInvalidateRegionCommand final { - explicit constexpr FlushAndInvalidateRegionCommand(VAddr addr, u64 size) - : addr{addr}, size{size} {} + explicit constexpr FlushAndInvalidateRegionCommand(VAddr addr_, u64 size_) + : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -92,8 +92,8 @@ using CommandData = struct CommandDataContainer { CommandDataContainer() = default; - CommandDataContainer(CommandData&& data, u64 next_fence) - : data{std::move(data)}, fence{next_fence} {} + explicit CommandDataContainer(CommandData&& data_, u64 next_fence_) + : data{std::move(data_)}, fence{next_fence_} {} CommandData data; u64 fence{}; @@ -112,7 +112,7 @@ struct SynchState final { /// Class used to manage the GPU thread class ThreadManager final { public: - explicit ThreadManager(Core::System& system); + explicit ThreadManager(Core::System& system_); ~ThreadManager(); /// Creates and starts the GPU thread. @@ -146,7 +146,6 @@ private: /// Pushes a command to be executed by the GPU thread u64 PushCommand(CommandData&& command_data); -private: SynchState state; Core::System& system; std::thread thread; diff --git a/src/video_core/guest_driver.h b/src/video_core/guest_driver.h index 99450777ee..21e569ba1f 100644 --- a/src/video_core/guest_driver.h +++ b/src/video_core/guest_driver.h @@ -19,8 +19,8 @@ namespace VideoCore { class GuestDriverProfile { public: explicit GuestDriverProfile() = default; - explicit GuestDriverProfile(std::optional texture_handler_size) - : texture_handler_size{texture_handler_size} {} + explicit GuestDriverProfile(std::optional texture_handler_size_) + : texture_handler_size{texture_handler_size_} {} void DeduceTextureHandlerSize(std::vector bound_offsets); diff --git a/src/video_core/macro/macro_hle.cpp b/src/video_core/macro/macro_hle.cpp index df00b57dff..70ac7c620b 100644 --- a/src/video_core/macro/macro_hle.cpp +++ b/src/video_core/macro/macro_hle.cpp @@ -85,7 +85,7 @@ constexpr std::array, 3> hle_funcs{{ {0x0217920100488FF7, &HLE_0217920100488FF7}, }}; -HLEMacro::HLEMacro(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {} +HLEMacro::HLEMacro(Engines::Maxwell3D& maxwell3d_) : maxwell3d{maxwell3d_} {} HLEMacro::~HLEMacro() = default; std::optional> HLEMacro::GetHLEProgram(u64 hash) const { @@ -99,8 +99,8 @@ std::optional> HLEMacro::GetHLEProgram(u64 hash) co HLEMacroImpl::~HLEMacroImpl() = default; -HLEMacroImpl::HLEMacroImpl(Engines::Maxwell3D& maxwell3d, HLEFunction func) - : maxwell3d(maxwell3d), func(func) {} +HLEMacroImpl::HLEMacroImpl(Engines::Maxwell3D& maxwell3d_, HLEFunction func_) + : maxwell3d{maxwell3d_}, func{func_} {} void HLEMacroImpl::Execute(const std::vector& parameters, u32 method) { func(maxwell3d, parameters); diff --git a/src/video_core/macro/macro_hle.h b/src/video_core/macro/macro_hle.h index 37af875a0e..cb3bd16009 100644 --- a/src/video_core/macro/macro_hle.h +++ b/src/video_core/macro/macro_hle.h @@ -20,7 +20,7 @@ using HLEFunction = void (*)(Engines::Maxwell3D& maxwell3d, const std::vector> GetHLEProgram(u64 hash) const; diff --git a/src/video_core/macro/macro_interpreter.cpp b/src/video_core/macro/macro_interpreter.cpp index bd01fd1f25..44a71aa6cd 100644 --- a/src/video_core/macro/macro_interpreter.cpp +++ b/src/video_core/macro/macro_interpreter.cpp @@ -11,29 +11,29 @@ MICROPROFILE_DEFINE(MacroInterp, "GPU", "Execute macro interpreter", MP_RGB(128, 128, 192)); namespace Tegra { -MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) - : MacroEngine::MacroEngine(maxwell3d), maxwell3d(maxwell3d) {} +MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d_) + : MacroEngine{maxwell3d_}, maxwell3d{maxwell3d_} {} std::unique_ptr MacroInterpreter::Compile(const std::vector& code) { return std::make_unique(maxwell3d, code); } -MacroInterpreterImpl::MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d, - const std::vector& code) - : maxwell3d(maxwell3d), code(code) {} +MacroInterpreterImpl::MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d_, + const std::vector& code_) + : maxwell3d{maxwell3d_}, code{code_} {} -void MacroInterpreterImpl::Execute(const std::vector& parameters, u32 method) { +void MacroInterpreterImpl::Execute(const std::vector& params, u32 method) { MICROPROFILE_SCOPE(MacroInterp); Reset(); - registers[1] = parameters[0]; - num_parameters = parameters.size(); + registers[1] = params[0]; + num_parameters = params.size(); if (num_parameters > parameters_capacity) { parameters_capacity = num_parameters; - this->parameters = std::make_unique(num_parameters); + parameters = std::make_unique(num_parameters); } - std::memcpy(this->parameters.get(), parameters.data(), num_parameters * sizeof(u32)); + std::memcpy(parameters.get(), params.data(), num_parameters * sizeof(u32)); // Execute the code until we hit an exit condition. bool keep_executing = true; diff --git a/src/video_core/macro/macro_interpreter.h b/src/video_core/macro/macro_interpreter.h index 90217fc89f..d50c619cee 100644 --- a/src/video_core/macro/macro_interpreter.h +++ b/src/video_core/macro/macro_interpreter.h @@ -17,7 +17,7 @@ class Maxwell3D; class MacroInterpreter final : public MacroEngine { public: - explicit MacroInterpreter(Engines::Maxwell3D& maxwell3d); + explicit MacroInterpreter(Engines::Maxwell3D& maxwell3d_); protected: std::unique_ptr Compile(const std::vector& code) override; @@ -28,8 +28,8 @@ private: class MacroInterpreterImpl : public CachedMacro { public: - MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d, const std::vector& code); - void Execute(const std::vector& parameters, u32 method) override; + explicit MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d_, const std::vector& code_); + void Execute(const std::vector& params, u32 method) override; private: /// Resets the execution engine state, zeroing registers, etc. @@ -38,9 +38,9 @@ private: /** * Executes a single macro instruction located at the current program counter. Returns whether * the interpreter should keep running. - * @param offset Offset to start execution at. + * * @param is_delay_slot Whether the current step is being executed due to a delay slot in a - * previous instruction. + * previous instruction. */ bool Step(bool is_delay_slot); diff --git a/src/video_core/macro/macro_jit_x64.cpp b/src/video_core/macro/macro_jit_x64.cpp index 954b87515a..c82bb987fd 100644 --- a/src/video_core/macro/macro_jit_x64.cpp +++ b/src/video_core/macro/macro_jit_x64.cpp @@ -28,15 +28,15 @@ static const std::bitset<32> PERSISTENT_REGISTERS = Common::X64::BuildRegSet({ BRANCH_HOLDER, }); -MacroJITx64::MacroJITx64(Engines::Maxwell3D& maxwell3d) - : MacroEngine::MacroEngine(maxwell3d), maxwell3d(maxwell3d) {} +MacroJITx64::MacroJITx64(Engines::Maxwell3D& maxwell3d_) + : MacroEngine{maxwell3d_}, maxwell3d{maxwell3d_} {} std::unique_ptr MacroJITx64::Compile(const std::vector& code) { return std::make_unique(maxwell3d, code); } -MacroJITx64Impl::MacroJITx64Impl(Engines::Maxwell3D& maxwell3d, const std::vector& code) - : Xbyak::CodeGenerator(MAX_CODE_SIZE), code(code), maxwell3d(maxwell3d) { +MacroJITx64Impl::MacroJITx64Impl(Engines::Maxwell3D& maxwell3d_, const std::vector& code_) + : CodeGenerator{MAX_CODE_SIZE}, code{code_}, maxwell3d{maxwell3d_} { Compile(); } @@ -553,15 +553,15 @@ Xbyak::Reg32 MacroJITx64Impl::Compile_GetRegister(u32 index, Xbyak::Reg32 dst) { } void MacroJITx64Impl::Compile_ProcessResult(Macro::ResultOperation operation, u32 reg) { - const auto SetRegister = [this](u32 reg, const Xbyak::Reg32& result) { + const auto SetRegister = [this](u32 reg_index, const Xbyak::Reg32& result) { // Register 0 is supposed to always return 0. NOP is implemented as a store to the zero // register. - if (reg == 0) { + if (reg_index == 0) { return; } - mov(dword[STATE + offsetof(JITState, registers) + reg * sizeof(u32)], result); + mov(dword[STATE + offsetof(JITState, registers) + reg_index * sizeof(u32)], result); }; - const auto SetMethodAddress = [this](const Xbyak::Reg32& reg) { mov(METHOD_ADDRESS, reg); }; + const auto SetMethodAddress = [this](const Xbyak::Reg32& reg32) { mov(METHOD_ADDRESS, reg32); }; switch (operation) { case Macro::ResultOperation::IgnoreAndFetch: diff --git a/src/video_core/macro/macro_jit_x64.h b/src/video_core/macro/macro_jit_x64.h index a180e74283..7f50ac2f88 100644 --- a/src/video_core/macro/macro_jit_x64.h +++ b/src/video_core/macro/macro_jit_x64.h @@ -23,7 +23,7 @@ constexpr size_t MAX_CODE_SIZE = 0x10000; class MacroJITx64 final : public MacroEngine { public: - explicit MacroJITx64(Engines::Maxwell3D& maxwell3d); + explicit MacroJITx64(Engines::Maxwell3D& maxwell3d_); protected: std::unique_ptr Compile(const std::vector& code) override; @@ -34,7 +34,7 @@ private: class MacroJITx64Impl : public Xbyak::CodeGenerator, public CachedMacro { public: - MacroJITx64Impl(Engines::Maxwell3D& maxwell3d, const std::vector& code); + explicit MacroJITx64Impl(Engines::Maxwell3D& maxwell3d_, const std::vector& code_); ~MacroJITx64Impl(); void Execute(const std::vector& parameters, u32 method) override; diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h index c078193d98..c35e576894 100644 --- a/src/video_core/memory_manager.h +++ b/src/video_core/memory_manager.h @@ -28,7 +28,7 @@ public: }; constexpr PageEntry() = default; - constexpr PageEntry(State state) : state{state} {} + constexpr PageEntry(State state_) : state{state_} {} constexpr PageEntry(VAddr addr) : state{static_cast(addr >> ShiftBits)} {} [[nodiscard]] constexpr bool IsUnmapped() const { @@ -68,7 +68,7 @@ static_assert(sizeof(PageEntry) == 4, "PageEntry is too large"); class MemoryManager final { public: - explicit MemoryManager(Core::System& system); + explicit MemoryManager(Core::System& system_); ~MemoryManager(); /// Binds a renderer to the memory manager. diff --git a/src/video_core/renderer_opengl/gl_arb_decompiler.cpp b/src/video_core/renderer_opengl/gl_arb_decompiler.cpp index d6120c23e0..5378c398e5 100644 --- a/src/video_core/renderer_opengl/gl_arb_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_arb_decompiler.cpp @@ -187,8 +187,8 @@ std::string TextureType(const MetaTexture& meta) { class ARBDecompiler final { public: - explicit ARBDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, - ShaderType stage, std::string_view identifier); + explicit ARBDecompiler(const Device& device_, const ShaderIR& ir_, const Registry& registry_, + ShaderType stage_, std::string_view identifier); std::string Code() const { return shader_source; @@ -802,9 +802,9 @@ private: }; }; -ARBDecompiler::ARBDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, - ShaderType stage, std::string_view identifier) - : device{device}, ir{ir}, registry{registry}, stage{stage} { +ARBDecompiler::ARBDecompiler(const Device& device_, const ShaderIR& ir_, const Registry& registry_, + ShaderType stage_, std::string_view identifier) + : device{device_}, ir{ir_}, registry{registry_}, stage{stage_} { DefineGlobalMemory(); AddLine("TEMP RC;"); @@ -1134,44 +1134,44 @@ void ARBDecompiler::VisitAST(const ASTNode& node) { for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { VisitAST(current); } - } else if (const auto ast = std::get_if(&*node->GetInnerData())) { - const std::string condition = VisitExpression(ast->condition); + } else if (const auto if_then = std::get_if(&*node->GetInnerData())) { + const std::string condition = VisitExpression(if_then->condition); ResetTemporaries(); AddLine("MOVC.U RC.x, {};", condition); AddLine("IF NE.x;"); - for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + for (ASTNode current = if_then->nodes.GetFirst(); current; current = current->GetNext()) { VisitAST(current); } AddLine("ENDIF;"); - } else if (const auto ast = std::get_if(&*node->GetInnerData())) { + } else if (const auto if_else = std::get_if(&*node->GetInnerData())) { AddLine("ELSE;"); - for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + for (ASTNode current = if_else->nodes.GetFirst(); current; current = current->GetNext()) { VisitAST(current); } - } else if (const auto ast = std::get_if(&*node->GetInnerData())) { - VisitBlock(ast->nodes); - } else if (const auto ast = std::get_if(&*node->GetInnerData())) { - AddLine("MOV.U F{}, {};", ast->index, VisitExpression(ast->condition)); + } else if (const auto decoded = std::get_if(&*node->GetInnerData())) { + VisitBlock(decoded->nodes); + } else if (const auto var_set = std::get_if(&*node->GetInnerData())) { + AddLine("MOV.U F{}, {};", var_set->index, VisitExpression(var_set->condition)); ResetTemporaries(); - } else if (const auto ast = std::get_if(&*node->GetInnerData())) { - const std::string condition = VisitExpression(ast->condition); + } else if (const auto do_while = std::get_if(&*node->GetInnerData())) { + const std::string condition = VisitExpression(do_while->condition); ResetTemporaries(); AddLine("REP;"); - for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + for (ASTNode current = do_while->nodes.GetFirst(); current; current = current->GetNext()) { VisitAST(current); } AddLine("MOVC.U RC.x, {};", condition); AddLine("BRK (NE.x);"); AddLine("ENDREP;"); - } else if (const auto ast = std::get_if(&*node->GetInnerData())) { - const bool is_true = ExprIsTrue(ast->condition); + } else if (const auto ast_return = std::get_if(&*node->GetInnerData())) { + const bool is_true = ExprIsTrue(ast_return->condition); if (!is_true) { - AddLine("MOVC.U RC.x, {};", VisitExpression(ast->condition)); + AddLine("MOVC.U RC.x, {};", VisitExpression(ast_return->condition)); AddLine("IF NE.x;"); ResetTemporaries(); } - if (ast->kills) { + if (ast_return->kills) { AddLine("KIL TR;"); } else { Exit(); @@ -1179,11 +1179,11 @@ void ARBDecompiler::VisitAST(const ASTNode& node) { if (!is_true) { AddLine("ENDIF;"); } - } else if (const auto ast = std::get_if(&*node->GetInnerData())) { - if (ExprIsTrue(ast->condition)) { + } else if (const auto ast_break = std::get_if(&*node->GetInnerData())) { + if (ExprIsTrue(ast_break->condition)) { AddLine("BRK;"); } else { - AddLine("MOVC.U RC.x, {};", VisitExpression(ast->condition)); + AddLine("MOVC.U RC.x, {};", VisitExpression(ast_break->condition)); AddLine("BRK (NE.x);"); ResetTemporaries(); } diff --git a/src/video_core/renderer_opengl/gl_fence_manager.cpp b/src/video_core/renderer_opengl/gl_fence_manager.cpp index b532fdcc25..6040646cbd 100644 --- a/src/video_core/renderer_opengl/gl_fence_manager.cpp +++ b/src/video_core/renderer_opengl/gl_fence_manager.cpp @@ -11,10 +11,10 @@ namespace OpenGL { -GLInnerFence::GLInnerFence(u32 payload, bool is_stubbed) : FenceBase(payload, is_stubbed) {} +GLInnerFence::GLInnerFence(u32 payload_, bool is_stubbed_) : FenceBase{payload_, is_stubbed_} {} -GLInnerFence::GLInnerFence(GPUVAddr address, u32 payload, bool is_stubbed) - : FenceBase(address, payload, is_stubbed) {} +GLInnerFence::GLInnerFence(GPUVAddr address_, u32 payload_, bool is_stubbed_) + : FenceBase{address_, payload_, is_stubbed_} {} GLInnerFence::~GLInnerFence() = default; @@ -45,10 +45,10 @@ void GLInnerFence::Wait() { glClientWaitSync(sync_object.handle, 0, GL_TIMEOUT_IGNORED); } -FenceManagerOpenGL::FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, - TextureCacheOpenGL& texture_cache, - OGLBufferCache& buffer_cache, QueryCache& query_cache) - : GenericFenceManager{rasterizer, gpu, texture_cache, buffer_cache, query_cache} {} +FenceManagerOpenGL::FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer_, + Tegra::GPU& gpu_, TextureCacheOpenGL& texture_cache_, + OGLBufferCache& buffer_cache_, QueryCache& query_cache_) + : GenericFenceManager{rasterizer_, gpu_, texture_cache_, buffer_cache_, query_cache_} {} Fence FenceManagerOpenGL::CreateFence(u32 value, bool is_stubbed) { return std::make_shared(value, is_stubbed); diff --git a/src/video_core/renderer_opengl/gl_fence_manager.h b/src/video_core/renderer_opengl/gl_fence_manager.h index da1dcdace9..39ca6125b6 100644 --- a/src/video_core/renderer_opengl/gl_fence_manager.h +++ b/src/video_core/renderer_opengl/gl_fence_manager.h @@ -17,8 +17,8 @@ namespace OpenGL { class GLInnerFence : public VideoCommon::FenceBase { public: - GLInnerFence(u32 payload, bool is_stubbed); - GLInnerFence(GPUVAddr address, u32 payload, bool is_stubbed); + explicit GLInnerFence(u32 payload_, bool is_stubbed_); + explicit GLInnerFence(GPUVAddr address_, u32 payload_, bool is_stubbed_); ~GLInnerFence(); void Queue(); @@ -37,9 +37,9 @@ using GenericFenceManager = class FenceManagerOpenGL final : public GenericFenceManager { public: - explicit FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, - TextureCacheOpenGL& texture_cache, OGLBufferCache& buffer_cache, - QueryCache& query_cache); + explicit FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_, + TextureCacheOpenGL& texture_cache_, OGLBufferCache& buffer_cache_, + QueryCache& query_cache_); protected: Fence CreateFence(u32 value, bool is_stubbed) override; diff --git a/src/video_core/renderer_opengl/gl_query_cache.cpp b/src/video_core/renderer_opengl/gl_query_cache.cpp index 1a3d9720ea..bcc37471fa 100644 --- a/src/video_core/renderer_opengl/gl_query_cache.cpp +++ b/src/video_core/renderer_opengl/gl_query_cache.cpp @@ -59,10 +59,10 @@ bool QueryCache::AnyCommandQueued() const noexcept { return gl_rasterizer.AnyCommandQueued(); } -HostCounter::HostCounter(QueryCache& cache, std::shared_ptr dependency, - VideoCore::QueryType type) - : VideoCommon::HostCounterBase{std::move(dependency)}, cache{cache}, - type{type}, query{cache.AllocateQuery(type)} { +HostCounter::HostCounter(QueryCache& cache_, std::shared_ptr dependency, + VideoCore::QueryType type_) + : HostCounterBase{std::move(dependency)}, cache{cache_}, type{type_}, + query{cache.AllocateQuery(type)} { glBeginQuery(GetTarget(type), query.handle); } @@ -86,13 +86,14 @@ u64 HostCounter::BlockingQuery() const { return static_cast(value); } -CachedQuery::CachedQuery(QueryCache& cache, VideoCore::QueryType type, VAddr cpu_addr, u8* host_ptr) - : VideoCommon::CachedQueryBase{cpu_addr, host_ptr}, cache{&cache}, type{type} {} +CachedQuery::CachedQuery(QueryCache& cache_, VideoCore::QueryType type_, VAddr cpu_addr, + u8* host_ptr) + : CachedQueryBase{cpu_addr, host_ptr}, cache{&cache_}, type{type_} {} CachedQuery::~CachedQuery() = default; CachedQuery::CachedQuery(CachedQuery&& rhs) noexcept - : VideoCommon::CachedQueryBase(std::move(rhs)), cache{rhs.cache}, type{rhs.type} {} + : CachedQueryBase(std::move(rhs)), cache{rhs.cache}, type{rhs.type} {} CachedQuery& CachedQuery::operator=(CachedQuery&& rhs) noexcept { cache = rhs.cache; diff --git a/src/video_core/renderer_opengl/gl_query_cache.h b/src/video_core/renderer_opengl/gl_query_cache.h index 82cac51ee5..d9851e880d 100644 --- a/src/video_core/renderer_opengl/gl_query_cache.h +++ b/src/video_core/renderer_opengl/gl_query_cache.h @@ -46,8 +46,8 @@ private: class HostCounter final : public VideoCommon::HostCounterBase { public: - explicit HostCounter(QueryCache& cache, std::shared_ptr dependency, - VideoCore::QueryType type); + explicit HostCounter(QueryCache& cache_, std::shared_ptr dependency, + VideoCore::QueryType type_); ~HostCounter(); void EndQuery(); @@ -62,7 +62,7 @@ private: class CachedQuery final : public VideoCommon::CachedQueryBase { public: - explicit CachedQuery(QueryCache& cache, VideoCore::QueryType type, VAddr cpu_addr, + explicit CachedQuery(QueryCache& cache_, VideoCore::QueryType type_, VAddr cpu_addr, u8* host_ptr); ~CachedQuery() override; diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index bd56bed0c6..9f2c0a2224 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -198,10 +198,10 @@ ProgramSharedPtr BuildShader(const Device& device, ShaderType shader_type, u64 u return program; } -Shader::Shader(std::shared_ptr registry_, ShaderEntries entries_, - ProgramSharedPtr program_, bool is_built) +Shader::Shader(std::shared_ptr registry_, ShaderEntries entries_, + ProgramSharedPtr program_, bool is_built_) : registry{std::move(registry_)}, entries{std::move(entries_)}, program{std::move(program_)}, - is_built(is_built) { + is_built{is_built_} { handle = program->assembly_program.handle; if (handle == 0) { handle = program->source_program.handle; diff --git a/src/video_core/renderer_opengl/gl_shader_cache.h b/src/video_core/renderer_opengl/gl_shader_cache.h index 1708af06af..ab5374fac6 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.h +++ b/src/video_core/renderer_opengl/gl_shader_cache.h @@ -108,7 +108,7 @@ public: private: explicit Shader(std::shared_ptr registry, ShaderEntries entries, - ProgramSharedPtr program, bool is_built = true); + ProgramSharedPtr program, bool is_built_ = true); std::shared_ptr registry; ShaderEntries entries; diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index 95ca96c8eb..0940969bac 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -131,7 +131,7 @@ private: class Expression final { public: - Expression(std::string code, Type type) : code{std::move(code)}, type{type} { + Expression(std::string code_, Type type_) : code{std::move(code_)}, type{type_} { ASSERT(type != Type::Void); } Expression() : type{Type::Void} {} @@ -148,8 +148,8 @@ public: ASSERT(type == Type::Void); } - std::string As(Type type) const { - switch (type) { + std::string As(Type type_) const { + switch (type_) { case Type::Bool: return AsBool(); case Type::Bool2: @@ -418,11 +418,12 @@ struct GenericVaryingDescription { class GLSLDecompiler final { public: - explicit GLSLDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, - ShaderType stage, std::string_view identifier, std::string_view suffix) - : device{device}, ir{ir}, registry{registry}, stage{stage}, identifier{identifier}, - suffix{suffix}, header{ir.GetHeader()}, use_unified_uniforms{ - UseUnifiedUniforms(device, ir, stage)} { + explicit GLSLDecompiler(const Device& device_, const ShaderIR& ir_, const Registry& registry_, + ShaderType stage_, std::string_view identifier_, + std::string_view suffix_) + : device{device_}, ir{ir_}, registry{registry_}, stage{stage_}, identifier{identifier_}, + suffix{suffix_}, header{ir.GetHeader()}, use_unified_uniforms{ + UseUnifiedUniforms(device_, ir_, stage_)} { if (stage != ShaderType::Compute) { transform_feedback = BuildTransformFeedback(registry.GetGraphicsInfo()); } @@ -777,16 +778,16 @@ private: name = "gs_" + name + "[]"; } - std::string suffix; + std::string suffix_; if (stage == ShaderType::Fragment) { const auto input_mode{header.ps.GetPixelImap(location)}; if (input_mode == PixelImap::Unused) { return; } - suffix = GetInputFlags(input_mode); + suffix_ = GetInputFlags(input_mode); } - code.AddLine("layout (location = {}) {} in vec4 {};", location, suffix, name); + code.AddLine("layout (location = {}) {} in vec4 {};", location, suffix_, name); } void DeclareOutputAttributes() { @@ -2100,13 +2101,13 @@ private: const auto type = meta.sampler.is_shadow ? Type::Float : Type::Int; const bool separate_dc = meta.sampler.is_shadow; - std::vector ir; + std::vector ir_; if (meta.sampler.is_shadow) { - ir = {TextureOffset{}}; + ir_ = {TextureOffset{}}; } else { - ir = {TextureOffset{}, TextureArgument{type, meta.component}}; + ir_ = {TextureOffset{}, TextureArgument{type, meta.component}}; } - return {GenerateTexture(operation, "Gather", ir, separate_dc) + GetSwizzle(meta.element), + return {GenerateTexture(operation, "Gather", ir_, separate_dc) + GetSwizzle(meta.element), Type::Float}; } @@ -2801,7 +2802,7 @@ std::string GetFlowVariable(u32 index) { class ExprDecompiler { public: - explicit ExprDecompiler(GLSLDecompiler& decomp) : decomp{decomp} {} + explicit ExprDecompiler(GLSLDecompiler& decomp_) : decomp{decomp_} {} void operator()(const ExprAnd& expr) { inner += '('; @@ -2856,7 +2857,7 @@ private: class ASTDecompiler { public: - explicit ASTDecompiler(GLSLDecompiler& decomp) : decomp{decomp} {} + explicit ASTDecompiler(GLSLDecompiler& decomp_) : decomp{decomp_} {} void operator()(const ASTProgram& ast) { ASTNode current = ast.nodes.GetFirst(); diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.h b/src/video_core/renderer_opengl/gl_shader_decompiler.h index 451c9689af..f5a5249f28 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.h +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.h @@ -25,8 +25,8 @@ using ImageEntry = VideoCommon::Shader::Image; class ConstBufferEntry : public VideoCommon::Shader::ConstBuffer { public: - explicit ConstBufferEntry(u32 max_offset, bool is_indirect, u32 index) - : VideoCommon::Shader::ConstBuffer{max_offset, is_indirect}, index{index} {} + explicit ConstBufferEntry(u32 max_offset, bool is_indirect, u32 index_) + : ConstBuffer{max_offset, is_indirect}, index{index_} {} u32 GetIndex() const { return index; @@ -37,10 +37,10 @@ private: }; struct GlobalMemoryEntry { - constexpr explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset, bool is_read, - bool is_written) - : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset}, is_read{is_read}, is_written{ - is_written} {} + constexpr explicit GlobalMemoryEntry(u32 cbuf_index_, u32 cbuf_offset_, bool is_read_, + bool is_written_) + : cbuf_index{cbuf_index_}, cbuf_offset{cbuf_offset_}, is_read{is_read_}, is_written{ + is_written_} {} u32 cbuf_index = 0; u32 cbuf_offset = 0; diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index a863ef2185..a59fe853ed 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -258,9 +258,9 @@ constexpr u32 EncodeSwizzle(SwizzleSource x_source, SwizzleSource y_source, Swiz } // Anonymous namespace -CachedSurface::CachedSurface(const GPUVAddr gpu_addr, const SurfaceParams& params, - bool is_astc_supported) - : VideoCommon::SurfaceBase(gpu_addr, params, is_astc_supported) { +CachedSurface::CachedSurface(const GPUVAddr gpu_addr_, const SurfaceParams& params_, + bool is_astc_supported_) + : SurfaceBase{gpu_addr_, params_, is_astc_supported_} { if (is_converted) { internal_format = params.srgb_conversion ? GL_SRGB8_ALPHA8 : GL_RGBA8; format = GL_RGBA; @@ -419,11 +419,11 @@ View CachedSurface::CreateViewInner(const ViewParams& view_key, const bool is_pr return view; } -CachedSurfaceView::CachedSurfaceView(CachedSurface& surface, const ViewParams& params, - bool is_proxy) - : VideoCommon::ViewBase(params), surface{surface}, format{surface.internal_format}, - target{GetTextureTarget(params.target)}, is_proxy{is_proxy} { - if (!is_proxy) { +CachedSurfaceView::CachedSurfaceView(CachedSurface& surface_, const ViewParams& params_, + bool is_proxy_) + : ViewBase{params_}, surface{surface_}, format{surface_.internal_format}, + target{GetTextureTarget(params_.target)}, is_proxy{is_proxy_} { + if (!is_proxy_) { main_view = CreateTextureView(); } } @@ -493,13 +493,13 @@ GLuint CachedSurfaceView::GetTexture(SwizzleSource x_source, SwizzleSource y_sou std::array swizzle{x_source, y_source, z_source, w_source}; - switch (const PixelFormat format = GetSurfaceParams().pixel_format) { + switch (const PixelFormat pixel_format = GetSurfaceParams().pixel_format) { case PixelFormat::D24_UNORM_S8_UINT: case PixelFormat::D32_FLOAT_S8_UINT: case PixelFormat::S8_UINT_D24_UNORM: UNIMPLEMENTED_IF(x_source != SwizzleSource::R && x_source != SwizzleSource::G); glTextureParameteri(view.handle, GL_DEPTH_STENCIL_TEXTURE_MODE, - GetComponent(format, x_source == SwizzleSource::R)); + GetComponent(pixel_format, x_source == SwizzleSource::R)); // Make sure we sample the first component std::transform(swizzle.begin(), swizzle.end(), swizzle.begin(), [](SwizzleSource value) { diff --git a/src/video_core/renderer_opengl/gl_texture_cache.h b/src/video_core/renderer_opengl/gl_texture_cache.h index 7787134fc6..76a7b23161 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.h +++ b/src/video_core/renderer_opengl/gl_texture_cache.h @@ -37,7 +37,8 @@ class CachedSurface final : public VideoCommon::SurfaceBase { friend CachedSurfaceView; public: - explicit CachedSurface(GPUVAddr gpu_addr, const SurfaceParams& params, bool is_astc_supported); + explicit CachedSurface(GPUVAddr gpu_addr_, const SurfaceParams& params_, + bool is_astc_supported_); ~CachedSurface(); void UploadTexture(const std::vector& staging_buffer) override; @@ -77,7 +78,7 @@ private: class CachedSurfaceView final : public VideoCommon::ViewBase { public: - explicit CachedSurfaceView(CachedSurface& surface, const ViewParams& params, bool is_proxy); + explicit CachedSurfaceView(CachedSurface& surface_, const ViewParams& params_, bool is_proxy_); ~CachedSurfaceView(); /// @brief Attaches this texture view to the currently bound fb_target framebuffer diff --git a/src/video_core/renderer_vulkan/vk_command_pool.cpp b/src/video_core/renderer_vulkan/vk_command_pool.cpp index 6339f4fe09..256a39148d 100644 --- a/src/video_core/renderer_vulkan/vk_command_pool.cpp +++ b/src/video_core/renderer_vulkan/vk_command_pool.cpp @@ -17,8 +17,8 @@ struct CommandPool::Pool { vk::CommandBuffers cmdbufs; }; -CommandPool::CommandPool(MasterSemaphore& master_semaphore, const VKDevice& device) - : ResourcePool(master_semaphore, COMMAND_BUFFER_POOL_SIZE), device{device} {} +CommandPool::CommandPool(MasterSemaphore& master_semaphore, const VKDevice& device_) + : ResourcePool(master_semaphore, COMMAND_BUFFER_POOL_SIZE), device{device_} {} CommandPool::~CommandPool() = default; diff --git a/src/video_core/renderer_vulkan/vk_command_pool.h b/src/video_core/renderer_vulkan/vk_command_pool.h index b9cb3fb5db..33655eca43 100644 --- a/src/video_core/renderer_vulkan/vk_command_pool.h +++ b/src/video_core/renderer_vulkan/vk_command_pool.h @@ -17,7 +17,7 @@ class VKDevice; class CommandPool final : public ResourcePool { public: - explicit CommandPool(MasterSemaphore& master_semaphore, const VKDevice& device); + explicit CommandPool(MasterSemaphore& master_semaphore, const VKDevice& device_); ~CommandPool() override; void Allocate(size_t begin, size_t end) override; -- cgit v1.2.3 From 341083a5e135edb06bcb8913f4897c232650d1da Mon Sep 17 00:00:00 2001 From: Lioncash Date: Sat, 5 Dec 2020 04:51:14 -0500 Subject: video_core: Resolve more variable shadowing scenarios pt.2 Migrates the video core code closer to enabling variable shadowing warnings as errors. This primarily sorts out shadowing occurrences within the Vulkan code. --- src/video_core/command_classes/codecs/vp9.cpp | 40 +++++++------- src/video_core/renderer_vulkan/renderer_vulkan.cpp | 4 +- src/video_core/renderer_vulkan/renderer_vulkan.h | 6 +-- src/video_core/renderer_vulkan/vk_compute_pass.cpp | 40 +++++++------- src/video_core/renderer_vulkan/vk_compute_pass.h | 23 ++++---- .../renderer_vulkan/vk_compute_pipeline.cpp | 16 +++--- .../renderer_vulkan/vk_compute_pipeline.h | 8 +-- src/video_core/renderer_vulkan/vk_device.cpp | 16 +++--- .../renderer_vulkan/vk_fence_manager.cpp | 22 ++++---- src/video_core/renderer_vulkan/vk_fence_manager.h | 16 +++--- .../renderer_vulkan/vk_graphics_pipeline.cpp | 38 +++++++------- .../renderer_vulkan/vk_graphics_pipeline.h | 14 ++--- src/video_core/renderer_vulkan/vk_image.cpp | 14 ++--- src/video_core/renderer_vulkan/vk_image.h | 4 +- .../renderer_vulkan/vk_memory_manager.cpp | 18 +++---- src/video_core/renderer_vulkan/vk_memory_manager.h | 10 ++-- src/video_core/renderer_vulkan/vk_query_cache.cpp | 30 +++++------ src/video_core/renderer_vulkan/vk_query_cache.h | 10 ++-- src/video_core/renderer_vulkan/vk_rasterizer.cpp | 17 +++--- src/video_core/renderer_vulkan/vk_rasterizer.h | 12 +++-- .../renderer_vulkan/vk_renderpass_cache.cpp | 2 +- .../renderer_vulkan/vk_renderpass_cache.h | 2 +- .../renderer_vulkan/vk_sampler_cache.cpp | 2 +- src/video_core/renderer_vulkan/vk_sampler_cache.h | 2 +- src/video_core/renderer_vulkan/vk_scheduler.h | 2 +- .../renderer_vulkan/vk_shader_decompiler.cpp | 28 +++++----- .../renderer_vulkan/vk_shader_decompiler.h | 8 +-- .../renderer_vulkan/vk_texture_cache.cpp | 61 +++++++++++----------- src/video_core/renderer_vulkan/vk_texture_cache.h | 18 +++---- .../renderer_vulkan/vk_update_descriptor.cpp | 4 +- .../renderer_vulkan/vk_update_descriptor.h | 2 +- src/video_core/renderer_vulkan/wrapper.cpp | 16 +++--- src/video_core/renderer_vulkan/wrapper.h | 25 +++++---- src/video_core/shader/control_flow.cpp | 6 +-- src/video_core/shader/decode/image.cpp | 10 ++-- src/video_core/shader/decode/other.cpp | 10 ++-- src/video_core/texture_cache/surface_base.cpp | 27 +++++----- src/video_core/texture_cache/surface_base.h | 4 +- src/video_core/texture_cache/surface_params.cpp | 14 ++--- 39 files changed, 305 insertions(+), 296 deletions(-) (limited to 'src/video_core/command_classes/codecs/vp9.cpp') diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp index 1771bc939b..59e586695a 100644 --- a/src/video_core/command_classes/codecs/vp9.cpp +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -374,43 +374,43 @@ void VP9::InsertEntropy(u64 offset, Vp9EntropyProbs& dst) { } Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) { - Vp9FrameContainer frame{}; + Vp9FrameContainer current_frame{}; { gpu.SyncGuestHost(); - frame.info = GetVp9PictureInfo(state); - frame.bit_stream.resize(frame.info.bitstream_size); - gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, frame.bit_stream.data(), - frame.info.bitstream_size); + current_frame.info = GetVp9PictureInfo(state); + current_frame.bit_stream.resize(current_frame.info.bitstream_size); + gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, current_frame.bit_stream.data(), + current_frame.info.bitstream_size); } // Buffer two frames, saving the last show frame info if (!next_next_frame.bit_stream.empty()) { Vp9FrameContainer temp{ - .info = frame.info, - .bit_stream = std::move(frame.bit_stream), + .info = current_frame.info, + .bit_stream = std::move(current_frame.bit_stream), }; - next_next_frame.info.show_frame = frame.info.last_frame_shown; - frame.info = next_next_frame.info; - frame.bit_stream = std::move(next_next_frame.bit_stream); + next_next_frame.info.show_frame = current_frame.info.last_frame_shown; + current_frame.info = next_next_frame.info; + current_frame.bit_stream = std::move(next_next_frame.bit_stream); next_next_frame = std::move(temp); if (!next_frame.bit_stream.empty()) { Vp9FrameContainer temp2{ - .info = frame.info, - .bit_stream = std::move(frame.bit_stream), + .info = current_frame.info, + .bit_stream = std::move(current_frame.bit_stream), }; - next_frame.info.show_frame = frame.info.last_frame_shown; - frame.info = next_frame.info; - frame.bit_stream = std::move(next_frame.bit_stream); + next_frame.info.show_frame = current_frame.info.last_frame_shown; + current_frame.info = next_frame.info; + current_frame.bit_stream = std::move(next_frame.bit_stream); next_frame = std::move(temp2); } else { - next_frame.info = frame.info; - next_frame.bit_stream = std::move(frame.bit_stream); + next_frame.info = current_frame.info; + next_frame.bit_stream = std::move(current_frame.bit_stream); } } else { - next_next_frame.info = frame.info; - next_next_frame.bit_stream = std::move(frame.bit_stream); + next_next_frame.info = current_frame.info; + next_next_frame.bit_stream = std::move(current_frame.bit_stream); } - return frame; + return current_frame; } std::vector VP9::ComposeCompressedHeader() { diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.cpp b/src/video_core/renderer_vulkan/renderer_vulkan.cpp index a2173edd27..ea4b7c1e64 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.cpp +++ b/src/video_core/renderer_vulkan/renderer_vulkan.cpp @@ -243,8 +243,8 @@ std::string BuildCommaSeparatedExtensions(std::vector available_ext RendererVulkan::RendererVulkan(Core::TelemetrySession& telemetry_session_, Core::Frontend::EmuWindow& emu_window, Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_, - std::unique_ptr context) - : RendererBase{emu_window, std::move(context)}, telemetry_session{telemetry_session_}, + std::unique_ptr context_) + : RendererBase{emu_window, std::move(context_)}, telemetry_session{telemetry_session_}, cpu_memory{cpu_memory_}, gpu{gpu_} {} RendererVulkan::~RendererVulkan() { diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.h b/src/video_core/renderer_vulkan/renderer_vulkan.h index 1044ca1244..977b86003f 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.h +++ b/src/video_core/renderer_vulkan/renderer_vulkan.h @@ -45,9 +45,9 @@ struct VKScreenInfo { class RendererVulkan final : public VideoCore::RendererBase { public: explicit RendererVulkan(Core::TelemetrySession& telemtry_session, - Core::Frontend::EmuWindow& emu_window, Core::Memory::Memory& cpu_memory, - Tegra::GPU& gpu, - std::unique_ptr context); + Core::Frontend::EmuWindow& emu_window, + Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_, + std::unique_ptr context_); ~RendererVulkan() override; bool Init() override; diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.cpp b/src/video_core/renderer_vulkan/vk_compute_pass.cpp index 9637c60596..1ac7e2a30c 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pass.cpp @@ -461,15 +461,15 @@ VkDescriptorSet VKComputePass::CommitDescriptorSet( return set; } -QuadArrayPass::QuadArrayPass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue) - : VKComputePass(device, descriptor_pool, BuildQuadArrayPassDescriptorSetLayoutBinding(), +QuadArrayPass::QuadArrayPass(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_) + : VKComputePass(device_, descriptor_pool_, BuildQuadArrayPassDescriptorSetLayoutBinding(), BuildQuadArrayPassDescriptorUpdateTemplateEntry(), BuildComputePushConstantRange(sizeof(u32)), std::size(quad_array), quad_array), - scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, - update_descriptor_queue{update_descriptor_queue} {} + scheduler{scheduler_}, staging_buffer_pool{staging_buffer_pool_}, + update_descriptor_queue{update_descriptor_queue_} {} QuadArrayPass::~QuadArrayPass() = default; @@ -510,14 +510,14 @@ std::pair QuadArrayPass::Assemble(u32 num_vertices, u32 return {*buffer.handle, 0}; } -Uint8Pass::Uint8Pass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue) - : VKComputePass(device, descriptor_pool, BuildInputOutputDescriptorSetBindings(), +Uint8Pass::Uint8Pass(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_) + : VKComputePass(device_, descriptor_pool_, BuildInputOutputDescriptorSetBindings(), BuildInputOutputDescriptorUpdateTemplate(), {}, std::size(uint8_pass), uint8_pass), - scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, - update_descriptor_queue{update_descriptor_queue} {} + scheduler{scheduler_}, staging_buffer_pool{staging_buffer_pool_}, + update_descriptor_queue{update_descriptor_queue_} {} Uint8Pass::~Uint8Pass() = default; @@ -555,16 +555,16 @@ std::pair Uint8Pass::Assemble(u32 num_vertices, VkBuffer src_buff return {*buffer.handle, 0}; } -QuadIndexedPass::QuadIndexedPass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue) - : VKComputePass(device, descriptor_pool, BuildInputOutputDescriptorSetBindings(), +QuadIndexedPass::QuadIndexedPass(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_) + : VKComputePass(device_, descriptor_pool_, BuildInputOutputDescriptorSetBindings(), BuildInputOutputDescriptorUpdateTemplate(), BuildComputePushConstantRange(sizeof(u32) * 2), std::size(QUAD_INDEXED_SPV), QUAD_INDEXED_SPV), - scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, - update_descriptor_queue{update_descriptor_queue} {} + scheduler{scheduler_}, staging_buffer_pool{staging_buffer_pool_}, + update_descriptor_queue{update_descriptor_queue_} {} QuadIndexedPass::~QuadIndexedPass() = default; diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.h b/src/video_core/renderer_vulkan/vk_compute_pass.h index acc94f27eb..2dc87902cf 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.h +++ b/src/video_core/renderer_vulkan/vk_compute_pass.h @@ -43,10 +43,10 @@ private: class QuadArrayPass final : public VKComputePass { public: - explicit QuadArrayPass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue); + explicit QuadArrayPass(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_); ~QuadArrayPass(); std::pair Assemble(u32 num_vertices, u32 first); @@ -59,9 +59,10 @@ private: class Uint8Pass final : public VKComputePass { public: - explicit Uint8Pass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue); + explicit Uint8Pass(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_); ~Uint8Pass(); std::pair Assemble(u32 num_vertices, VkBuffer src_buffer, u64 src_offset); @@ -74,10 +75,10 @@ private: class QuadIndexedPass final : public VKComputePass { public: - explicit QuadIndexedPass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue); + explicit QuadIndexedPass(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_); ~QuadIndexedPass(); std::pair Assemble(Tegra::Engines::Maxwell3D::Regs::IndexFormat index_format, diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp index 9be72dc9b6..62f44d6da2 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp @@ -15,16 +15,16 @@ namespace Vulkan { -VKComputePipeline::VKComputePipeline(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - const SPIRVShader& shader) - : device{device}, scheduler{scheduler}, entries{shader.entries}, +VKComputePipeline::VKComputePipeline(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + const SPIRVShader& shader_) + : device{device_}, scheduler{scheduler_}, entries{shader_.entries}, descriptor_set_layout{CreateDescriptorSetLayout()}, - descriptor_allocator{descriptor_pool, *descriptor_set_layout}, - update_descriptor_queue{update_descriptor_queue}, layout{CreatePipelineLayout()}, + descriptor_allocator{descriptor_pool_, *descriptor_set_layout}, + update_descriptor_queue{update_descriptor_queue_}, layout{CreatePipelineLayout()}, descriptor_template{CreateDescriptorUpdateTemplate()}, - shader_module{CreateShaderModule(shader.code)}, pipeline{CreatePipeline()} {} + shader_module{CreateShaderModule(shader_.code)}, pipeline{CreatePipeline()} {} VKComputePipeline::~VKComputePipeline() = default; diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.h b/src/video_core/renderer_vulkan/vk_compute_pipeline.h index 6e2f22a4a7..49e2113a2c 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.h +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.h @@ -17,10 +17,10 @@ class VKUpdateDescriptorQueue; class VKComputePipeline final { public: - explicit VKComputePipeline(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - const SPIRVShader& shader); + explicit VKComputePipeline(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + const SPIRVShader& shader_); ~VKComputePipeline(); VkDescriptorSet CommitDescriptorSet(); diff --git a/src/video_core/renderer_vulkan/vk_device.cpp b/src/video_core/renderer_vulkan/vk_device.cpp index f34ed6735c..ce3846195c 100644 --- a/src/video_core/renderer_vulkan/vk_device.cpp +++ b/src/video_core/renderer_vulkan/vk_device.cpp @@ -491,8 +491,8 @@ bool VKDevice::IsOptimalAstcSupported(const VkPhysicalDeviceFeatures& features) VK_FORMAT_FEATURE_BLIT_DST_BIT | VK_FORMAT_FEATURE_TRANSFER_SRC_BIT | VK_FORMAT_FEATURE_TRANSFER_DST_BIT}; for (const auto format : astc_formats) { - const auto format_properties{physical.GetFormatProperties(format)}; - if (!(format_properties.optimalTilingFeatures & format_feature_usage)) { + const auto physical_format_properties{physical.GetFormatProperties(format)}; + if ((physical_format_properties.optimalTilingFeatures & format_feature_usage) == 0) { return false; } } @@ -644,8 +644,8 @@ std::vector VKDevice::LoadExtensions() { VkPhysicalDeviceFeatures2KHR features; features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FEATURES_2_KHR; - VkPhysicalDeviceProperties2KHR properties; - properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR; + VkPhysicalDeviceProperties2KHR physical_properties; + physical_properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR; if (has_khr_shader_float16_int8) { VkPhysicalDeviceFloat16Int8FeaturesKHR float16_int8_features; @@ -670,8 +670,8 @@ std::vector VKDevice::LoadExtensions() { subgroup_properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SUBGROUP_SIZE_CONTROL_PROPERTIES_EXT; subgroup_properties.pNext = nullptr; - properties.pNext = &subgroup_properties; - physical.GetProperties2KHR(properties); + physical_properties.pNext = &subgroup_properties; + physical.GetProperties2KHR(physical_properties); is_warp_potentially_bigger = subgroup_properties.maxSubgroupSize > GuestWarpSize; @@ -695,8 +695,8 @@ std::vector VKDevice::LoadExtensions() { VkPhysicalDeviceTransformFeedbackPropertiesEXT tfb_properties; tfb_properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TRANSFORM_FEEDBACK_PROPERTIES_EXT; tfb_properties.pNext = nullptr; - properties.pNext = &tfb_properties; - physical.GetProperties2KHR(properties); + physical_properties.pNext = &tfb_properties; + physical.GetProperties2KHR(physical_properties); if (tfb_features.transformFeedback && tfb_features.geometryStreams && tfb_properties.maxTransformFeedbackStreams >= 4 && diff --git a/src/video_core/renderer_vulkan/vk_fence_manager.cpp b/src/video_core/renderer_vulkan/vk_fence_manager.cpp index 5babbdd0b7..0bcaee7147 100644 --- a/src/video_core/renderer_vulkan/vk_fence_manager.cpp +++ b/src/video_core/renderer_vulkan/vk_fence_manager.cpp @@ -14,12 +14,13 @@ namespace Vulkan { -InnerFence::InnerFence(const VKDevice& device, VKScheduler& scheduler, u32 payload, bool is_stubbed) - : VideoCommon::FenceBase(payload, is_stubbed), device{device}, scheduler{scheduler} {} +InnerFence::InnerFence(const VKDevice& device_, VKScheduler& scheduler_, u32 payload_, + bool is_stubbed_) + : FenceBase{payload_, is_stubbed_}, device{device_}, scheduler{scheduler_} {} -InnerFence::InnerFence(const VKDevice& device, VKScheduler& scheduler, GPUVAddr address, - u32 payload, bool is_stubbed) - : VideoCommon::FenceBase(address, payload, is_stubbed), device{device}, scheduler{scheduler} {} +InnerFence::InnerFence(const VKDevice& device_, VKScheduler& scheduler_, GPUVAddr address_, + u32 payload_, bool is_stubbed_) + : FenceBase{address_, payload_, is_stubbed_}, device{device_}, scheduler{scheduler_} {} InnerFence::~InnerFence() = default; @@ -71,11 +72,12 @@ bool InnerFence::IsEventSignalled() const { } } -VKFenceManager::VKFenceManager(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, - Tegra::MemoryManager& memory_manager, VKTextureCache& texture_cache, - VKBufferCache& buffer_cache, VKQueryCache& query_cache, - const VKDevice& device_, VKScheduler& scheduler_) - : GenericFenceManager(rasterizer, gpu, texture_cache, buffer_cache, query_cache), +VKFenceManager::VKFenceManager(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_, + Tegra::MemoryManager& memory_manager_, + VKTextureCache& texture_cache_, VKBufferCache& buffer_cache_, + VKQueryCache& query_cache_, const VKDevice& device_, + VKScheduler& scheduler_) + : GenericFenceManager{rasterizer_, gpu_, texture_cache_, buffer_cache_, query_cache_}, device{device_}, scheduler{scheduler_} {} Fence VKFenceManager::CreateFence(u32 value, bool is_stubbed) { diff --git a/src/video_core/renderer_vulkan/vk_fence_manager.h b/src/video_core/renderer_vulkan/vk_fence_manager.h index 1547d6d30d..c8547cc24e 100644 --- a/src/video_core/renderer_vulkan/vk_fence_manager.h +++ b/src/video_core/renderer_vulkan/vk_fence_manager.h @@ -28,10 +28,10 @@ class VKTextureCache; class InnerFence : public VideoCommon::FenceBase { public: - explicit InnerFence(const VKDevice& device, VKScheduler& scheduler, u32 payload, - bool is_stubbed); - explicit InnerFence(const VKDevice& device, VKScheduler& scheduler, GPUVAddr address, - u32 payload, bool is_stubbed); + explicit InnerFence(const VKDevice& device_, VKScheduler& scheduler_, u32 payload_, + bool is_stubbed_); + explicit InnerFence(const VKDevice& device_, VKScheduler& scheduler_, GPUVAddr address_, + u32 payload_, bool is_stubbed_); ~InnerFence(); void Queue(); @@ -55,10 +55,10 @@ using GenericFenceManager = class VKFenceManager final : public GenericFenceManager { public: - explicit VKFenceManager(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, - Tegra::MemoryManager& memory_manager, VKTextureCache& texture_cache, - VKBufferCache& buffer_cache, VKQueryCache& query_cache, - const VKDevice& device, VKScheduler& scheduler); + explicit VKFenceManager(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_, + Tegra::MemoryManager& memory_manager_, VKTextureCache& texture_cache_, + VKBufferCache& buffer_cache_, VKQueryCache& query_cache_, + const VKDevice& device_, VKScheduler& scheduler_); protected: Fence CreateFence(u32 value, bool is_stubbed) override; diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp index 0e8f9c352d..f8a1bcf34b 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp @@ -71,21 +71,21 @@ VkViewportSwizzleNV UnpackViewportSwizzle(u16 swizzle) { } // Anonymous namespace -VKGraphicsPipeline::VKGraphicsPipeline(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - VKRenderPassCache& renderpass_cache, - const GraphicsPipelineCacheKey& key, - vk::Span bindings, - const SPIRVProgram& program) - : device{device}, scheduler{scheduler}, cache_key{key}, hash{cache_key.Hash()}, - descriptor_set_layout{CreateDescriptorSetLayout(bindings)}, - descriptor_allocator{descriptor_pool, *descriptor_set_layout}, - update_descriptor_queue{update_descriptor_queue}, layout{CreatePipelineLayout()}, - descriptor_template{CreateDescriptorUpdateTemplate(program)}, modules{CreateShaderModules( - program)}, - renderpass{renderpass_cache.GetRenderPass(cache_key.renderpass_params)}, - pipeline{CreatePipeline(cache_key.renderpass_params, program)} {} +VKGraphicsPipeline::VKGraphicsPipeline(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + VKRenderPassCache& renderpass_cache_, + const GraphicsPipelineCacheKey& key_, + vk::Span bindings_, + const SPIRVProgram& program_) + : device{device_}, scheduler{scheduler_}, cache_key{key_}, hash{cache_key.Hash()}, + descriptor_set_layout{CreateDescriptorSetLayout(bindings_)}, + descriptor_allocator{descriptor_pool_, *descriptor_set_layout}, + update_descriptor_queue{update_descriptor_queue_}, layout{CreatePipelineLayout()}, + descriptor_template{CreateDescriptorUpdateTemplate(program_)}, modules{CreateShaderModules( + program_)}, + renderpass{renderpass_cache_.GetRenderPass(cache_key.renderpass_params)}, + pipeline{CreatePipeline(cache_key.renderpass_params, program_)} {} VKGraphicsPipeline::~VKGraphicsPipeline() = default; @@ -162,8 +162,8 @@ std::vector VKGraphicsPipeline::CreateShaderModules( .codeSize = 0, }; - std::vector modules; - modules.reserve(Maxwell::MaxShaderStage); + std::vector shader_modules; + shader_modules.reserve(Maxwell::MaxShaderStage); for (std::size_t i = 0; i < Maxwell::MaxShaderStage; ++i) { const auto& stage = program[i]; if (!stage) { @@ -174,9 +174,9 @@ std::vector VKGraphicsPipeline::CreateShaderModules( ci.codeSize = stage->code.size() * sizeof(u32); ci.pCode = stage->code.data(); - modules.push_back(device.GetLogical().CreateShaderModule(ci)); + shader_modules.push_back(device.GetLogical().CreateShaderModule(ci)); } - return modules; + return shader_modules; } vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpass_params, diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.h b/src/video_core/renderer_vulkan/vk_graphics_pipeline.h index 58aa35efdd..3fb31d55ab 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.h +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.h @@ -51,13 +51,13 @@ using SPIRVProgram = std::array, Maxwell::MaxShaderSt class VKGraphicsPipeline final { public: - explicit VKGraphicsPipeline(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - VKRenderPassCache& renderpass_cache, - const GraphicsPipelineCacheKey& key, - vk::Span bindings, - const SPIRVProgram& program); + explicit VKGraphicsPipeline(const VKDevice& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + VKRenderPassCache& renderpass_cache_, + const GraphicsPipelineCacheKey& key_, + vk::Span bindings_, + const SPIRVProgram& program_); ~VKGraphicsPipeline(); VkDescriptorSet CommitDescriptorSet(); diff --git a/src/video_core/renderer_vulkan/vk_image.cpp b/src/video_core/renderer_vulkan/vk_image.cpp index 1c418ea17d..072d14e3b0 100644 --- a/src/video_core/renderer_vulkan/vk_image.cpp +++ b/src/video_core/renderer_vulkan/vk_image.cpp @@ -13,18 +13,18 @@ namespace Vulkan { -VKImage::VKImage(const VKDevice& device, VKScheduler& scheduler, const VkImageCreateInfo& image_ci, - VkImageAspectFlags aspect_mask) - : device{device}, scheduler{scheduler}, format{image_ci.format}, aspect_mask{aspect_mask}, - image_num_layers{image_ci.arrayLayers}, image_num_levels{image_ci.mipLevels} { - UNIMPLEMENTED_IF_MSG(image_ci.queueFamilyIndexCount != 0, +VKImage::VKImage(const VKDevice& device_, VKScheduler& scheduler_, + const VkImageCreateInfo& image_ci_, VkImageAspectFlags aspect_mask_) + : device{device_}, scheduler{scheduler_}, format{image_ci_.format}, aspect_mask{aspect_mask_}, + image_num_layers{image_ci_.arrayLayers}, image_num_levels{image_ci_.mipLevels} { + UNIMPLEMENTED_IF_MSG(image_ci_.queueFamilyIndexCount != 0, "Queue family tracking is not implemented"); - image = device.GetLogical().CreateImage(image_ci); + image = device_.GetLogical().CreateImage(image_ci_); const u32 num_ranges = image_num_layers * image_num_levels; barriers.resize(num_ranges); - subrange_states.resize(num_ranges, {{}, image_ci.initialLayout}); + subrange_states.resize(num_ranges, {{}, image_ci_.initialLayout}); } VKImage::~VKImage() = default; diff --git a/src/video_core/renderer_vulkan/vk_image.h b/src/video_core/renderer_vulkan/vk_image.h index b4d7229e54..287ab90ca0 100644 --- a/src/video_core/renderer_vulkan/vk_image.h +++ b/src/video_core/renderer_vulkan/vk_image.h @@ -17,8 +17,8 @@ class VKScheduler; class VKImage { public: - explicit VKImage(const VKDevice& device, VKScheduler& scheduler, - const VkImageCreateInfo& image_ci, VkImageAspectFlags aspect_mask); + explicit VKImage(const VKDevice& device_, VKScheduler& scheduler_, + const VkImageCreateInfo& image_ci_, VkImageAspectFlags aspect_mask_); ~VKImage(); /// Records in the passed command buffer an image transition and updates the state of the image. diff --git a/src/video_core/renderer_vulkan/vk_memory_manager.cpp b/src/video_core/renderer_vulkan/vk_memory_manager.cpp index 24c8960acc..be53d450fa 100644 --- a/src/video_core/renderer_vulkan/vk_memory_manager.cpp +++ b/src/video_core/renderer_vulkan/vk_memory_manager.cpp @@ -29,10 +29,10 @@ u64 GetAllocationChunkSize(u64 required_size) { class VKMemoryAllocation final { public: - explicit VKMemoryAllocation(const VKDevice& device, vk::DeviceMemory memory, - VkMemoryPropertyFlags properties, u64 allocation_size, u32 type) - : device{device}, memory{std::move(memory)}, properties{properties}, - allocation_size{allocation_size}, shifted_type{ShiftType(type)} {} + explicit VKMemoryAllocation(const VKDevice& device_, vk::DeviceMemory memory_, + VkMemoryPropertyFlags properties_, u64 allocation_size_, u32 type_) + : device{device_}, memory{std::move(memory_)}, properties{properties_}, + allocation_size{allocation_size_}, shifted_type{ShiftType(type_)} {} VKMemoryCommit Commit(VkDeviceSize commit_size, VkDeviceSize alignment) { auto found = TryFindFreeSection(free_iterator, allocation_size, @@ -117,8 +117,8 @@ private: std::vector commits; }; -VKMemoryManager::VKMemoryManager(const VKDevice& device) - : device{device}, properties{device.GetPhysical().GetMemoryProperties()} {} +VKMemoryManager::VKMemoryManager(const VKDevice& device_) + : device{device_}, properties{device_.GetPhysical().GetMemoryProperties()} {} VKMemoryManager::~VKMemoryManager() = default; @@ -207,9 +207,9 @@ VKMemoryCommit VKMemoryManager::TryAllocCommit(const VkMemoryRequirements& requi return {}; } -VKMemoryCommitImpl::VKMemoryCommitImpl(const VKDevice& device, VKMemoryAllocation* allocation, - const vk::DeviceMemory& memory, u64 begin, u64 end) - : device{device}, memory{memory}, interval{begin, end}, allocation{allocation} {} +VKMemoryCommitImpl::VKMemoryCommitImpl(const VKDevice& device_, VKMemoryAllocation* allocation_, + const vk::DeviceMemory& memory_, u64 begin_, u64 end_) + : device{device_}, memory{memory_}, interval{begin_, end_}, allocation{allocation_} {} VKMemoryCommitImpl::~VKMemoryCommitImpl() { allocation->Free(this); diff --git a/src/video_core/renderer_vulkan/vk_memory_manager.h b/src/video_core/renderer_vulkan/vk_memory_manager.h index 1af88e3d4b..39f903ec85 100644 --- a/src/video_core/renderer_vulkan/vk_memory_manager.h +++ b/src/video_core/renderer_vulkan/vk_memory_manager.h @@ -21,7 +21,7 @@ using VKMemoryCommit = std::unique_ptr; class VKMemoryManager final { public: - explicit VKMemoryManager(const VKDevice& device); + explicit VKMemoryManager(const VKDevice& device_); VKMemoryManager(const VKMemoryManager&) = delete; ~VKMemoryManager(); @@ -58,8 +58,8 @@ class VKMemoryCommitImpl final { friend MemoryMap; public: - explicit VKMemoryCommitImpl(const VKDevice& device, VKMemoryAllocation* allocation, - const vk::DeviceMemory& memory, u64 begin, u64 end); + explicit VKMemoryCommitImpl(const VKDevice& device_, VKMemoryAllocation* allocation_, + const vk::DeviceMemory& memory_, u64 begin_, u64 end_); ~VKMemoryCommitImpl(); /// Maps a memory region and returns a pointer to it. @@ -93,8 +93,8 @@ private: /// Holds ownership of a memory map. class MemoryMap final { public: - explicit MemoryMap(const VKMemoryCommitImpl* commit, u8* address) - : commit{commit}, address{address} {} + explicit MemoryMap(const VKMemoryCommitImpl* commit_, u8* address_) + : commit{commit_}, address{address_} {} ~MemoryMap() { if (commit) { diff --git a/src/video_core/renderer_vulkan/vk_query_cache.cpp b/src/video_core/renderer_vulkan/vk_query_cache.cpp index ee2d871e31..6fa0717370 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_query_cache.cpp @@ -66,15 +66,15 @@ void QueryPool::Reserve(std::pair query) { usage[pool_index * GROW_STEP + static_cast(query.second)] = false; } -VKQueryCache::VKQueryCache(VideoCore::RasterizerInterface& rasterizer, - Tegra::Engines::Maxwell3D& maxwell3d, Tegra::MemoryManager& gpu_memory, - const VKDevice& device, VKScheduler& scheduler) - : VideoCommon::QueryCacheBase{rasterizer, maxwell3d, gpu_memory}, - device{device}, scheduler{scheduler}, query_pools{ - QueryPool{device, scheduler, - QueryType::SamplesPassed}, - } {} +VKQueryCache::VKQueryCache(VideoCore::RasterizerInterface& rasterizer_, + Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::MemoryManager& gpu_memory_, + const VKDevice& device_, VKScheduler& scheduler_) + : QueryCacheBase{rasterizer_, maxwell3d_, + gpu_memory_}, + device{device_}, scheduler{scheduler_}, query_pools{ + QueryPool{device_, scheduler_, + QueryType::SamplesPassed}, + } {} VKQueryCache::~VKQueryCache() { // TODO(Rodrigo): This is a hack to destroy all HostCounter instances before the base class @@ -95,12 +95,12 @@ void VKQueryCache::Reserve(QueryType type, std::pair query) { query_pools[static_cast(type)].Reserve(query); } -HostCounter::HostCounter(VKQueryCache& cache, std::shared_ptr dependency, - QueryType type) - : VideoCommon::HostCounterBase{std::move(dependency)}, cache{cache}, - type{type}, query{cache.AllocateQuery(type)}, tick{cache.Scheduler().CurrentTick()} { - const vk::Device* logical = &cache.Device().GetLogical(); - cache.Scheduler().Record([logical, query = query](vk::CommandBuffer cmdbuf) { +HostCounter::HostCounter(VKQueryCache& cache_, std::shared_ptr dependency_, + QueryType type_) + : HostCounterBase{std::move(dependency_)}, cache{cache_}, + type{type_}, query{cache_.AllocateQuery(type_)}, tick{cache_.Scheduler().CurrentTick()} { + const vk::Device* logical = &cache_.Device().GetLogical(); + cache_.Scheduler().Record([logical, query = query](vk::CommandBuffer cmdbuf) { logical->ResetQueryPoolEXT(query.first, query.second, 1); cmdbuf.BeginQuery(query.first, query.second, VK_QUERY_CONTROL_PRECISE_BIT); }); diff --git a/src/video_core/renderer_vulkan/vk_query_cache.h b/src/video_core/renderer_vulkan/vk_query_cache.h index 2e57fb75d4..201fca8885 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.h +++ b/src/video_core/renderer_vulkan/vk_query_cache.h @@ -53,9 +53,9 @@ private: class VKQueryCache final : public VideoCommon::QueryCacheBase { public: - explicit VKQueryCache(VideoCore::RasterizerInterface& rasterizer, - Tegra::Engines::Maxwell3D& maxwell3d, Tegra::MemoryManager& gpu_memory, - const VKDevice& device, VKScheduler& scheduler); + explicit VKQueryCache(VideoCore::RasterizerInterface& rasterizer_, + Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::MemoryManager& gpu_memory_, + const VKDevice& device_, VKScheduler& scheduler_); ~VKQueryCache(); std::pair AllocateQuery(VideoCore::QueryType type); @@ -78,8 +78,8 @@ private: class HostCounter final : public VideoCommon::HostCounterBase { public: - explicit HostCounter(VKQueryCache& cache, std::shared_ptr dependency, - VideoCore::QueryType type); + explicit HostCounter(VKQueryCache& cache_, std::shared_ptr dependency_, + VideoCore::QueryType type_); ~HostCounter(); void EndQuery(); diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index e0fb8693f6..560386081d 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -904,15 +904,14 @@ void RasterizerVulkan::SetupShaderDescriptors( texture_cache.GuardSamplers(false); } -void RasterizerVulkan::SetupImageTransitions( - Texceptions texceptions, const std::array& color_attachments, - const View& zeta_attachment) { +void RasterizerVulkan::SetupImageTransitions(Texceptions texceptions, const ColorAttachments& color, + const ZetaAttachment& zeta) { TransitionImages(sampled_views, VK_PIPELINE_STAGE_ALL_GRAPHICS_BIT, VK_ACCESS_SHADER_READ_BIT); TransitionImages(image_views, VK_PIPELINE_STAGE_ALL_GRAPHICS_BIT, VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT); - for (std::size_t rt = 0; rt < std::size(color_attachments); ++rt) { - const auto color_attachment = color_attachments[rt]; + for (std::size_t rt = 0; rt < color.size(); ++rt) { + const auto color_attachment = color[rt]; if (color_attachment == nullptr) { continue; } @@ -923,13 +922,13 @@ void RasterizerVulkan::SetupImageTransitions( VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT); } - if (zeta_attachment != nullptr) { + if (zeta != nullptr) { const auto image_layout = texceptions[ZETA_TEXCEPTION_INDEX] ? VK_IMAGE_LAYOUT_GENERAL : VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL; - zeta_attachment->Transition(image_layout, VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT, - VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | - VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT); + zeta->Transition(image_layout, VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT, + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT); } } diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.h b/src/video_core/renderer_vulkan/vk_rasterizer.h index 237e51fa4e..1789fb2850 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.h +++ b/src/video_core/renderer_vulkan/vk_rasterizer.h @@ -160,6 +160,9 @@ private: bool is_indexed = 0; }; + using ColorAttachments = std::array; + using ZetaAttachment = View; + using Texceptions = std::bitset; static constexpr std::size_t ZETA_TEXCEPTION_INDEX = 8; @@ -181,9 +184,8 @@ private: /// Setup descriptors in the graphics pipeline. void SetupShaderDescriptors(const std::array& shaders); - void SetupImageTransitions(Texceptions texceptions, - const std::array& color_attachments, - const View& zeta_attachment); + void SetupImageTransitions(Texceptions texceptions, const ColorAttachments& color, + const ZetaAttachment& zeta); void UpdateDynamicStates(); @@ -308,8 +310,8 @@ private: vk::Event wfi_event; VideoCommon::Shader::AsyncShaders async_shaders; - std::array color_attachments; - View zeta_attachment; + ColorAttachments color_attachments; + ZetaAttachment zeta_attachment; std::vector sampled_views; std::vector image_views; diff --git a/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp b/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp index 80284cf92c..e812c7dd6a 100644 --- a/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp @@ -24,7 +24,7 @@ bool RenderPassParams::operator==(const RenderPassParams& rhs) const noexcept { return std::memcmp(&rhs, this, sizeof *this) == 0; } -VKRenderPassCache::VKRenderPassCache(const VKDevice& device) : device{device} {} +VKRenderPassCache::VKRenderPassCache(const VKDevice& device_) : device{device_} {} VKRenderPassCache::~VKRenderPassCache() = default; diff --git a/src/video_core/renderer_vulkan/vk_renderpass_cache.h b/src/video_core/renderer_vulkan/vk_renderpass_cache.h index 8b0fec7203..652ecef7b2 100644 --- a/src/video_core/renderer_vulkan/vk_renderpass_cache.h +++ b/src/video_core/renderer_vulkan/vk_renderpass_cache.h @@ -55,7 +55,7 @@ namespace Vulkan { class VKRenderPassCache final { public: - explicit VKRenderPassCache(const VKDevice& device); + explicit VKRenderPassCache(const VKDevice& device_); ~VKRenderPassCache(); VkRenderPass GetRenderPass(const RenderPassParams& params); diff --git a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp index b068888f95..b859691fa7 100644 --- a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp @@ -36,7 +36,7 @@ VkBorderColor ConvertBorderColor(std::array color) { } // Anonymous namespace -VKSamplerCache::VKSamplerCache(const VKDevice& device) : device{device} {} +VKSamplerCache::VKSamplerCache(const VKDevice& device_) : device{device_} {} VKSamplerCache::~VKSamplerCache() = default; diff --git a/src/video_core/renderer_vulkan/vk_sampler_cache.h b/src/video_core/renderer_vulkan/vk_sampler_cache.h index a33d1c0ee6..3f22c46104 100644 --- a/src/video_core/renderer_vulkan/vk_sampler_cache.h +++ b/src/video_core/renderer_vulkan/vk_sampler_cache.h @@ -14,7 +14,7 @@ class VKDevice; class VKSamplerCache final : public VideoCommon::SamplerCache { public: - explicit VKSamplerCache(const VKDevice& device); + explicit VKSamplerCache(const VKDevice& device_); ~VKSamplerCache(); protected: diff --git a/src/video_core/renderer_vulkan/vk_scheduler.h b/src/video_core/renderer_vulkan/vk_scheduler.h index 7be8a19f07..6d3a5da0bc 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.h +++ b/src/video_core/renderer_vulkan/vk_scheduler.h @@ -104,7 +104,7 @@ private: template class TypedCommand final : public Command { public: - explicit TypedCommand(T&& command) : command{std::move(command)} {} + explicit TypedCommand(T&& command_) : command{std::move(command_)} {} ~TypedCommand() override = default; TypedCommand(TypedCommand&&) = delete; diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp index fed9ebecda..7b0169acd7 100644 --- a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp @@ -55,8 +55,8 @@ enum class Type { Void, Bool, Bool2, Float, Int, Uint, HalfFloat }; class Expression final { public: - Expression(Id id, Type type) : id{id}, type{type} { - ASSERT(type != Type::Void); + Expression(Id id_, Type type_) : id{id_}, type{type_} { + ASSERT(type_ != Type::Void); } Expression() : type{Type::Void} {} @@ -281,12 +281,12 @@ u32 ShaderVersion(const VKDevice& device) { class SPIRVDecompiler final : public Sirit::Module { public: - explicit SPIRVDecompiler(const VKDevice& device, const ShaderIR& ir, ShaderType stage, - const Registry& registry, const Specialization& specialization) - : Module(ShaderVersion(device)), device{device}, ir{ir}, stage{stage}, - header{ir.GetHeader()}, registry{registry}, specialization{specialization} { - if (stage != ShaderType::Compute) { - transform_feedback = BuildTransformFeedback(registry.GetGraphicsInfo()); + explicit SPIRVDecompiler(const VKDevice& device_, const ShaderIR& ir_, ShaderType stage_, + const Registry& registry_, const Specialization& specialization_) + : Module(ShaderVersion(device_)), device{device_}, ir{ir_}, stage{stage_}, + header{ir_.GetHeader()}, registry{registry_}, specialization{specialization_} { + if (stage_ != ShaderType::Compute) { + transform_feedback = BuildTransformFeedback(registry_.GetGraphicsInfo()); } AddCapability(spv::Capability::Shader); @@ -330,7 +330,7 @@ public: if (device.IsFloat16Supported()) { AddCapability(spv::Capability::Float16); } - t_scalar_half = Name(TypeFloat(device.IsFloat16Supported() ? 16 : 32), "scalar_half"); + t_scalar_half = Name(TypeFloat(device_.IsFloat16Supported() ? 16 : 32), "scalar_half"); t_half = Name(TypeVector(t_scalar_half, 2), "half"); const Id main = Decompile(); @@ -1088,9 +1088,9 @@ private: indices.point_size = AddBuiltIn(t_float, spv::BuiltIn::PointSize, "point_size"); } - const auto& output_attributes = ir.GetOutputAttributes(); - const bool declare_clip_distances = - std::any_of(output_attributes.begin(), output_attributes.end(), [](const auto& index) { + const auto& ir_output_attributes = ir.GetOutputAttributes(); + const bool declare_clip_distances = std::any_of( + ir_output_attributes.begin(), ir_output_attributes.end(), [](const auto& index) { return index == Attribute::Index::ClipDistances0123 || index == Attribute::Index::ClipDistances4567; }); @@ -2891,7 +2891,7 @@ private: class ExprDecompiler { public: - explicit ExprDecompiler(SPIRVDecompiler& decomp) : decomp{decomp} {} + explicit ExprDecompiler(SPIRVDecompiler& decomp_) : decomp{decomp_} {} Id operator()(const ExprAnd& expr) { const Id type_def = decomp.GetTypeDefinition(Type::Bool); @@ -2947,7 +2947,7 @@ private: class ASTDecompiler { public: - explicit ASTDecompiler(SPIRVDecompiler& decomp) : decomp{decomp} {} + explicit ASTDecompiler(SPIRVDecompiler& decomp_) : decomp{decomp_} {} void operator()(const ASTProgram& ast) { ASTNode current = ast.nodes.GetFirst(); diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.h b/src/video_core/renderer_vulkan/vk_shader_decompiler.h index 1108489228..df18125147 100644 --- a/src/video_core/renderer_vulkan/vk_shader_decompiler.h +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.h @@ -30,8 +30,8 @@ constexpr u32 DESCRIPTOR_SET = 0; class ConstBufferEntry : public VideoCommon::Shader::ConstBuffer { public: - explicit constexpr ConstBufferEntry(const VideoCommon::Shader::ConstBuffer& entry, u32 index) - : VideoCommon::Shader::ConstBuffer{entry}, index{index} {} + explicit constexpr ConstBufferEntry(const ConstBuffer& entry_, u32 index_) + : ConstBuffer{entry_}, index{index_} {} constexpr u32 GetIndex() const { return index; @@ -43,8 +43,8 @@ private: class GlobalBufferEntry { public: - constexpr explicit GlobalBufferEntry(u32 cbuf_index, u32 cbuf_offset, bool is_written) - : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset}, is_written{is_written} {} + constexpr explicit GlobalBufferEntry(u32 cbuf_index_, u32 cbuf_offset_, bool is_written_) + : cbuf_index{cbuf_index_}, cbuf_offset{cbuf_offset_}, is_written{is_written_} {} constexpr u32 GetCbufIndex() const { return cbuf_index; diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index f2c8f2ae19..64649699fd 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -180,19 +180,19 @@ VkImageCreateInfo GenerateImageCreateInfo(const VKDevice& device, const SurfaceP return ci; } -u32 EncodeSwizzle(Tegra::Texture::SwizzleSource x_source, Tegra::Texture::SwizzleSource y_source, - Tegra::Texture::SwizzleSource z_source, Tegra::Texture::SwizzleSource w_source) { +u32 EncodeSwizzle(SwizzleSource x_source, SwizzleSource y_source, SwizzleSource z_source, + SwizzleSource w_source) { return (static_cast(x_source) << 24) | (static_cast(y_source) << 16) | (static_cast(z_source) << 8) | static_cast(w_source); } } // Anonymous namespace -CachedSurface::CachedSurface(const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler, VKStagingBufferPool& staging_pool, - GPUVAddr gpu_addr, const SurfaceParams& params) - : SurfaceBase{gpu_addr, params, device.IsOptimalAstcSupported()}, device{device}, - memory_manager{memory_manager}, scheduler{scheduler}, staging_pool{staging_pool} { +CachedSurface::CachedSurface(const VKDevice& device_, VKMemoryManager& memory_manager_, + VKScheduler& scheduler_, VKStagingBufferPool& staging_pool_, + GPUVAddr gpu_addr_, const SurfaceParams& params_) + : SurfaceBase{gpu_addr_, params_, device_.IsOptimalAstcSupported()}, device{device_}, + memory_manager{memory_manager_}, scheduler{scheduler_}, staging_pool{staging_pool_} { if (params.IsBuffer()) { buffer = CreateBuffer(device, params, host_memory_size); commit = memory_manager.Commit(buffer, false); @@ -234,7 +234,7 @@ void CachedSurface::UploadTexture(const std::vector& staging_buffer) { void CachedSurface::DownloadTexture(std::vector& staging_buffer) { UNIMPLEMENTED_IF(params.IsBuffer()); - if (params.pixel_format == VideoCore::Surface::PixelFormat::A1B5G5R5_UNORM) { + if (params.pixel_format == PixelFormat::A1B5G5R5_UNORM) { LOG_WARNING(Render_Vulkan, "A1B5G5R5 flushing is stubbed"); } @@ -244,10 +244,10 @@ void CachedSurface::DownloadTexture(std::vector& staging_buffer) { FullTransition(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_ACCESS_TRANSFER_READ_BIT, VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL); - const auto& buffer = staging_pool.GetUnusedBuffer(host_memory_size, true); + const auto& unused_buffer = staging_pool.GetUnusedBuffer(host_memory_size, true); // TODO(Rodrigo): Do this in a single copy for (u32 level = 0; level < params.num_levels; ++level) { - scheduler.Record([image = *image->GetHandle(), buffer = *buffer.handle, + scheduler.Record([image = *image->GetHandle(), buffer = *unused_buffer.handle, copy = GetBufferImageCopy(level)](vk::CommandBuffer cmdbuf) { cmdbuf.CopyImageToBuffer(image, VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, buffer, copy); }); @@ -255,16 +255,17 @@ void CachedSurface::DownloadTexture(std::vector& staging_buffer) { scheduler.Finish(); // TODO(Rodrigo): Use an intern buffer for staging buffers and avoid this unnecessary memcpy. - std::memcpy(staging_buffer.data(), buffer.commit->Map(host_memory_size), host_memory_size); + std::memcpy(staging_buffer.data(), unused_buffer.commit->Map(host_memory_size), + host_memory_size); } void CachedSurface::DecorateSurfaceName() { // TODO(Rodrigo): Add name decorations } -View CachedSurface::CreateView(const ViewParams& params) { +View CachedSurface::CreateView(const ViewParams& view_params) { // TODO(Rodrigo): Add name decorations - return views[params] = std::make_shared(device, *this, params); + return views[view_params] = std::make_shared(device, *this, view_params); } void CachedSurface::UploadBuffer(const std::vector& staging_buffer) { @@ -348,21 +349,21 @@ VkImageSubresourceRange CachedSurface::GetImageSubresourceRange() const { static_cast(params.GetNumLayers())}; } -CachedSurfaceView::CachedSurfaceView(const VKDevice& device, CachedSurface& surface, - const ViewParams& params) - : VideoCommon::ViewBase{params}, params{surface.GetSurfaceParams()}, - image{surface.GetImageHandle()}, buffer_view{surface.GetBufferViewHandle()}, - aspect_mask{surface.GetAspectMask()}, device{device}, surface{surface}, - base_level{params.base_level}, num_levels{params.num_levels}, - image_view_type{image ? GetImageViewType(params.target) : VK_IMAGE_VIEW_TYPE_1D} { +CachedSurfaceView::CachedSurfaceView(const VKDevice& device_, CachedSurface& surface_, + const ViewParams& view_params_) + : ViewBase{view_params_}, surface_params{surface_.GetSurfaceParams()}, + image{surface_.GetImageHandle()}, buffer_view{surface_.GetBufferViewHandle()}, + aspect_mask{surface_.GetAspectMask()}, device{device_}, surface{surface_}, + base_level{view_params_.base_level}, num_levels{view_params_.num_levels}, + image_view_type{image ? GetImageViewType(view_params_.target) : VK_IMAGE_VIEW_TYPE_1D} { if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) { base_layer = 0; num_layers = 1; - base_slice = params.base_layer; - num_slices = params.num_layers; + base_slice = view_params_.base_layer; + num_slices = view_params_.num_layers; } else { - base_layer = params.base_layer; - num_layers = params.num_layers; + base_layer = view_params_.base_layer; + num_layers = view_params_.num_layers; } } @@ -384,7 +385,7 @@ VkImageView CachedSurfaceView::GetImageView(SwizzleSource x_source, SwizzleSourc std::array swizzle{MaxwellToVK::SwizzleSource(x_source), MaxwellToVK::SwizzleSource(y_source), MaxwellToVK::SwizzleSource(z_source), MaxwellToVK::SwizzleSource(w_source)}; - if (params.pixel_format == VideoCore::Surface::PixelFormat::A1B5G5R5_UNORM) { + if (surface_params.pixel_format == PixelFormat::A1B5G5R5_UNORM) { // A1B5G5R5 is implemented as A1R5G5B5, we have to change the swizzle here. std::swap(swizzle[0], swizzle[2]); } @@ -395,12 +396,12 @@ VkImageView CachedSurfaceView::GetImageView(SwizzleSource x_source, SwizzleSourc if (aspect == (VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT)) { UNIMPLEMENTED_IF(x_source != SwizzleSource::R && x_source != SwizzleSource::G); const bool is_first = x_source == SwizzleSource::R; - switch (params.pixel_format) { - case VideoCore::Surface::PixelFormat::D24_UNORM_S8_UINT: - case VideoCore::Surface::PixelFormat::D32_FLOAT_S8_UINT: + switch (surface_params.pixel_format) { + case PixelFormat::D24_UNORM_S8_UINT: + case PixelFormat::D32_FLOAT_S8_UINT: aspect = is_first ? VK_IMAGE_ASPECT_DEPTH_BIT : VK_IMAGE_ASPECT_STENCIL_BIT; break; - case VideoCore::Surface::PixelFormat::S8_UINT_D24_UNORM: + case PixelFormat::S8_UINT_D24_UNORM: aspect = is_first ? VK_IMAGE_ASPECT_STENCIL_BIT : VK_IMAGE_ASPECT_DEPTH_BIT; break; default: @@ -417,7 +418,7 @@ VkImageView CachedSurfaceView::GetImageView(SwizzleSource x_source, SwizzleSourc if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) { ASSERT(base_slice == 0); - ASSERT(num_slices == params.depth); + ASSERT(num_slices == surface_params.depth); } image_view = device.GetLogical().CreateImageView({ diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h index 39202feba2..06880f2285 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.h +++ b/src/video_core/renderer_vulkan/vk_texture_cache.h @@ -40,9 +40,9 @@ class CachedSurface final : public VideoCommon::SurfaceBase { friend CachedSurfaceView; public: - explicit CachedSurface(const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler, VKStagingBufferPool& staging_pool, - GPUVAddr gpu_addr, const SurfaceParams& params); + explicit CachedSurface(const VKDevice& device_, VKMemoryManager& memory_manager_, + VKScheduler& scheduler_, VKStagingBufferPool& staging_pool_, + GPUVAddr gpu_addr_, const SurfaceParams& params_); ~CachedSurface(); void UploadTexture(const std::vector& staging_buffer) override; @@ -84,7 +84,7 @@ public: protected: void DecorateSurfaceName(); - View CreateView(const ViewParams& params) override; + View CreateView(const ViewParams& view_params) override; private: void UploadBuffer(const std::vector& staging_buffer); @@ -110,8 +110,8 @@ private: class CachedSurfaceView final : public VideoCommon::ViewBase { public: - explicit CachedSurfaceView(const VKDevice& device, CachedSurface& surface, - const ViewParams& params); + explicit CachedSurfaceView(const VKDevice& device_, CachedSurface& surface_, + const ViewParams& view_params_); ~CachedSurfaceView(); VkImageView GetImageView(Tegra::Texture::SwizzleSource x_source, @@ -126,11 +126,11 @@ public: } u32 GetWidth() const { - return params.GetMipWidth(base_level); + return surface_params.GetMipWidth(base_level); } u32 GetHeight() const { - return params.GetMipHeight(base_level); + return surface_params.GetMipHeight(base_level); } u32 GetNumLayers() const { @@ -169,7 +169,7 @@ public: private: // Store a copy of these values to avoid double dereference when reading them - const SurfaceParams params; + const SurfaceParams surface_params; const VkImage image; const VkBufferView buffer_view; const VkImageAspectFlags aspect_mask; diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp index 351c048d23..8826da3252 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp @@ -14,8 +14,8 @@ namespace Vulkan { -VKUpdateDescriptorQueue::VKUpdateDescriptorQueue(const VKDevice& device, VKScheduler& scheduler) - : device{device}, scheduler{scheduler} {} +VKUpdateDescriptorQueue::VKUpdateDescriptorQueue(const VKDevice& device_, VKScheduler& scheduler_) + : device{device_}, scheduler{scheduler_} {} VKUpdateDescriptorQueue::~VKUpdateDescriptorQueue() = default; diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.h b/src/video_core/renderer_vulkan/vk_update_descriptor.h index 945320c724..f7e3c98211 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.h +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.h @@ -31,7 +31,7 @@ struct DescriptorUpdateEntry { class VKUpdateDescriptorQueue final { public: - explicit VKUpdateDescriptorQueue(const VKDevice& device, VKScheduler& scheduler); + explicit VKUpdateDescriptorQueue(const VKDevice& device_, VKScheduler& scheduler_); ~VKUpdateDescriptorQueue(); void TickFrame(); diff --git a/src/video_core/renderer_vulkan/wrapper.cpp b/src/video_core/renderer_vulkan/wrapper.cpp index 4e83303d89..1eced809ee 100644 --- a/src/video_core/renderer_vulkan/wrapper.cpp +++ b/src/video_core/renderer_vulkan/wrapper.cpp @@ -417,7 +417,7 @@ VkResult Free(VkDevice device, VkCommandPool handle, Span buffe } Instance Instance::Create(u32 version, Span layers, Span extensions, - InstanceDispatch& dld) noexcept { + InstanceDispatch& dispatch) noexcept { const VkApplicationInfo application_info{ .sType = VK_STRUCTURE_TYPE_APPLICATION_INFO, .pNext = nullptr, @@ -439,17 +439,17 @@ Instance Instance::Create(u32 version, Span layers, Span> Instance::EnumeratePhysicalDevices() { @@ -540,7 +540,7 @@ std::vector SwapchainKHR::GetImages() const { Device Device::Create(VkPhysicalDevice physical_device, Span queues_ci, Span enabled_extensions, const void* next, - DeviceDispatch& dld) noexcept { + DeviceDispatch& dispatch) noexcept { const VkDeviceCreateInfo ci{ .sType = VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO, .pNext = next, @@ -555,11 +555,11 @@ Device Device::Create(VkPhysicalDevice physical_device, Span allocations, std::size_t num, - VkDevice device, PoolType pool, const DeviceDispatch& dld) noexcept - : allocations{std::move(allocations)}, num{num}, device{device}, pool{pool}, dld{&dld} {} + explicit PoolAllocations(std::unique_ptr allocations_, std::size_t num_, + VkDevice device_, PoolType pool_, const DeviceDispatch& dld_) noexcept + : allocations{std::move(allocations_)}, num{num_}, device{device_}, pool{pool_}, + dld{&dld_} {} /// Copying Vulkan allocations is not supported and will never be. PoolAllocations(const PoolAllocations&) = delete; @@ -565,7 +566,7 @@ class Instance : public Handle { public: /// Creates a Vulkan instance. Use "operator bool" for error handling. static Instance Create(u32 version, Span layers, Span extensions, - InstanceDispatch& dld) noexcept; + InstanceDispatch& dispatch) noexcept; /// Enumerates physical devices. /// @return Physical devices and an empty handle on failure. @@ -581,7 +582,8 @@ public: constexpr Queue() noexcept = default; /// Construct a queue handle. - constexpr Queue(VkQueue queue, const DeviceDispatch& dld) noexcept : queue{queue}, dld{&dld} {} + constexpr Queue(VkQueue queue_, const DeviceDispatch& dld_) noexcept + : queue{queue_}, dld{&dld_} {} VkResult Submit(Span submit_infos, VkFence fence = VK_NULL_HANDLE) const noexcept { @@ -720,7 +722,7 @@ class Device : public Handle { public: static Device Create(VkPhysicalDevice physical_device, Span queues_ci, Span enabled_extensions, const void* next, - DeviceDispatch& dld) noexcept; + DeviceDispatch& dispatch) noexcept; Queue GetQueue(u32 family_index) const noexcept; @@ -809,8 +811,9 @@ class PhysicalDevice { public: constexpr PhysicalDevice() noexcept = default; - constexpr PhysicalDevice(VkPhysicalDevice physical_device, const InstanceDispatch& dld) noexcept - : physical_device{physical_device}, dld{&dld} {} + constexpr PhysicalDevice(VkPhysicalDevice physical_device_, + const InstanceDispatch& dld_) noexcept + : physical_device{physical_device_}, dld{&dld_} {} constexpr operator VkPhysicalDevice() const noexcept { return physical_device; @@ -849,8 +852,8 @@ class CommandBuffer { public: CommandBuffer() noexcept = default; - explicit CommandBuffer(VkCommandBuffer handle, const DeviceDispatch& dld) noexcept - : handle{handle}, dld{&dld} {} + explicit CommandBuffer(VkCommandBuffer handle_, const DeviceDispatch& dld_) noexcept + : handle{handle_}, dld{&dld_} {} const VkCommandBuffer* address() const noexcept { return &handle; diff --git a/src/video_core/shader/control_flow.cpp b/src/video_core/shader/control_flow.cpp index 4c8971615b..d656e0668c 100644 --- a/src/video_core/shader/control_flow.cpp +++ b/src/video_core/shader/control_flow.cpp @@ -241,10 +241,10 @@ std::pair ParseCode(CFGRebuildState& state, u32 address) ParseInfo parse_info{}; SingleBranch single_branch{}; - const auto insert_label = [](CFGRebuildState& state, u32 address) { - const auto pair = state.labels.emplace(address); + const auto insert_label = [](CFGRebuildState& rebuild_state, u32 label_address) { + const auto pair = rebuild_state.labels.emplace(label_address); if (pair.second) { - state.inspect_queries.push_back(address); + rebuild_state.inspect_queries.push_back(label_address); } }; diff --git a/src/video_core/shader/decode/image.cpp b/src/video_core/shader/decode/image.cpp index 1ed4212eee..532f66d27d 100644 --- a/src/video_core/shader/decode/image.cpp +++ b/src/video_core/shader/decode/image.cpp @@ -358,9 +358,9 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { instr.suldst.GetStoreDataLayout() != StoreType::Bits64); auto descriptor = [this, instr] { - std::optional descriptor; + std::optional sampler_descriptor; if (instr.suldst.is_immediate) { - descriptor = + sampler_descriptor = registry.ObtainBoundSampler(static_cast(instr.image.index.Value())); } else { const Node image_register = GetRegister(instr.gpr39); @@ -368,12 +368,12 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { static_cast(global_code.size())); const auto buffer = std::get<1>(result); const auto offset = std::get<2>(result); - descriptor = registry.ObtainBindlessSampler(buffer, offset); + sampler_descriptor = registry.ObtainBindlessSampler(buffer, offset); } - if (!descriptor) { + if (!sampler_descriptor) { UNREACHABLE_MSG("Failed to obtain image descriptor"); } - return *descriptor; + return *sampler_descriptor; }(); const auto comp_mask = GetImageComponentMask(descriptor.format); diff --git a/src/video_core/shader/decode/other.cpp b/src/video_core/shader/decode/other.cpp index 29a7cfbfe4..1db500bc45 100644 --- a/src/video_core/shader/decode/other.cpp +++ b/src/video_core/shader/decode/other.cpp @@ -90,11 +90,11 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { UNIMPLEMENTED_MSG("S2R WscaleFactorZ is not implemented"); return Immediate(0U); case SystemVariable::Tid: { - Node value = Immediate(0); - value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdX), 0, 9); - value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdY), 16, 9); - value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdZ), 26, 5); - return value; + Node val = Immediate(0); + val = BitfieldInsert(val, Operation(OperationCode::LocalInvocationIdX), 0, 9); + val = BitfieldInsert(val, Operation(OperationCode::LocalInvocationIdY), 16, 9); + val = BitfieldInsert(val, Operation(OperationCode::LocalInvocationIdZ), 26, 5); + return val; } case SystemVariable::TidX: return Operation(OperationCode::LocalInvocationIdX); diff --git a/src/video_core/texture_cache/surface_base.cpp b/src/video_core/texture_cache/surface_base.cpp index b44c09d710..42a1c0c6f2 100644 --- a/src/video_core/texture_cache/surface_base.cpp +++ b/src/video_core/texture_cache/surface_base.cpp @@ -167,27 +167,28 @@ std::vector SurfaceBaseImpl::BreakDownNonLayered(const SurfaceParams return result; } -void SurfaceBaseImpl::SwizzleFunc(MortonSwizzleMode mode, u8* memory, const SurfaceParams& params, - u8* buffer, u32 level) { - const u32 width{params.GetMipWidth(level)}; - const u32 height{params.GetMipHeight(level)}; - const u32 block_height{params.GetMipBlockHeight(level)}; - const u32 block_depth{params.GetMipBlockDepth(level)}; +void SurfaceBaseImpl::SwizzleFunc(MortonSwizzleMode mode, u8* memory, + const SurfaceParams& surface_params, u8* buffer, u32 level) { + const u32 width{surface_params.GetMipWidth(level)}; + const u32 height{surface_params.GetMipHeight(level)}; + const u32 block_height{surface_params.GetMipBlockHeight(level)}; + const u32 block_depth{surface_params.GetMipBlockDepth(level)}; std::size_t guest_offset{mipmap_offsets[level]}; - if (params.is_layered) { + if (surface_params.is_layered) { std::size_t host_offset = 0; const std::size_t guest_stride = layer_size; - const std::size_t host_stride = params.GetHostLayerSize(level); - for (u32 layer = 0; layer < params.depth; ++layer) { - MortonSwizzle(mode, params.pixel_format, width, block_height, height, block_depth, 1, - params.tile_width_spacing, buffer + host_offset, memory + guest_offset); + const std::size_t host_stride = surface_params.GetHostLayerSize(level); + for (u32 layer = 0; layer < surface_params.depth; ++layer) { + MortonSwizzle(mode, surface_params.pixel_format, width, block_height, height, + block_depth, 1, surface_params.tile_width_spacing, buffer + host_offset, + memory + guest_offset); guest_offset += guest_stride; host_offset += host_stride; } } else { - MortonSwizzle(mode, params.pixel_format, width, block_height, height, block_depth, - params.GetMipDepth(level), params.tile_width_spacing, buffer, + MortonSwizzle(mode, surface_params.pixel_format, width, block_height, height, block_depth, + surface_params.GetMipDepth(level), surface_params.tile_width_spacing, buffer, memory + guest_offset); } } diff --git a/src/video_core/texture_cache/surface_base.h b/src/video_core/texture_cache/surface_base.h index 173f2edba0..cfcfa5b3a7 100644 --- a/src/video_core/texture_cache/surface_base.h +++ b/src/video_core/texture_cache/surface_base.h @@ -167,8 +167,8 @@ protected: std::vector mipmap_offsets; private: - void SwizzleFunc(MortonSwizzleMode mode, u8* memory, const SurfaceParams& params, u8* buffer, - u32 level); + void SwizzleFunc(MortonSwizzleMode mode, u8* memory, const SurfaceParams& surface_params, + u8* buffer, u32 level); std::vector BreakDownLayered(const SurfaceParams& in_params) const; diff --git a/src/video_core/texture_cache/surface_params.cpp b/src/video_core/texture_cache/surface_params.cpp index 13dd163563..3052977196 100644 --- a/src/video_core/texture_cache/surface_params.cpp +++ b/src/video_core/texture_cache/surface_params.cpp @@ -356,18 +356,18 @@ std::size_t SurfaceParams::GetLayerSize(bool as_host_size, bool uncompressed) co std::size_t SurfaceParams::GetInnerMipmapMemorySize(u32 level, bool as_host_size, bool uncompressed) const { - const u32 width{GetMipmapSize(uncompressed, GetMipWidth(level), GetDefaultBlockWidth())}; - const u32 height{GetMipmapSize(uncompressed, GetMipHeight(level), GetDefaultBlockHeight())}; - const u32 depth{is_layered ? 1U : GetMipDepth(level)}; + const u32 mip_width{GetMipmapSize(uncompressed, GetMipWidth(level), GetDefaultBlockWidth())}; + const u32 mip_height{GetMipmapSize(uncompressed, GetMipHeight(level), GetDefaultBlockHeight())}; + const u32 mip_depth{is_layered ? 1U : GetMipDepth(level)}; if (is_tiled) { - return Tegra::Texture::CalculateSize(!as_host_size, GetBytesPerPixel(), width, height, - depth, GetMipBlockHeight(level), + return Tegra::Texture::CalculateSize(!as_host_size, GetBytesPerPixel(), mip_width, + mip_height, mip_depth, GetMipBlockHeight(level), GetMipBlockDepth(level)); } else if (as_host_size || IsBuffer()) { - return GetBytesPerPixel() * width * height * depth; + return GetBytesPerPixel() * mip_width * mip_height * mip_depth; } else { // Linear Texture Case - return pitch * height * depth; + return pitch * mip_height * mip_depth; } } -- cgit v1.2.3