From 5da70f719703084482933e103e561cc98163f370 Mon Sep 17 00:00:00 2001 From: Kelebek1 Date: Tue, 23 May 2023 14:45:54 +0100 Subject: [PATCH] Remove memory allocations in some hot paths --- src/audio_core/device/audio_buffers.h | 8 ++- src/audio_core/device/device_session.cpp | 12 ++-- src/audio_core/device/device_session.h | 7 ++- src/audio_core/in/audio_in_system.cpp | 5 +- src/audio_core/out/audio_out_system.cpp | 4 +- .../renderer/command/data_source/decode.cpp | 23 ++++---- .../renderer/command/effect/compressor.cpp | 8 +-- .../renderer/command/effect/delay.cpp | 14 ++--- .../renderer/command/effect/i3dl2_reverb.cpp | 4 +- .../renderer/command/effect/light_limiter.cpp | 12 ++-- .../renderer/command/effect/reverb.cpp | 12 ++-- .../renderer/command/sink/circular_buffer.cpp | 4 +- .../renderer/command/sink/device.cpp | 5 +- src/audio_core/renderer/mix/mix_context.cpp | 6 +- src/audio_core/renderer/nodes/node_states.cpp | 4 +- src/audio_core/renderer/nodes/node_states.h | 2 +- src/audio_core/renderer/system.cpp | 1 + src/audio_core/sink/null_sink.h | 2 +- src/audio_core/sink/sink_stream.cpp | 15 ++--- src/audio_core/sink/sink_stream.h | 5 +- src/common/ring_buffer.h | 3 +- src/common/scratch_buffer.h | 9 +++ .../hle/kernel/k_synchronization_object.cpp | 3 +- src/core/hle/kernel/k_thread.cpp | 8 +-- src/core/hle/kernel/k_thread.h | 3 +- src/core/hle/kernel/svc/svc_ipc.cpp | 7 ++- .../hle/kernel/svc/svc_synchronization.cpp | 10 ++-- src/core/hle/kernel/svc/svc_thread.cpp | 2 +- src/core/hle/service/audio/audin_u.cpp | 16 ++--- src/core/hle/service/audio/audout_u.cpp | 15 ++--- src/core/hle/service/audio/audren_u.cpp | 22 +++---- src/core/hle/service/audio/audren_u.h | 1 + src/core/hle/service/audio/hwopus.cpp | 9 +-- src/core/hle/service/nvdrv/devices/nvdevice.h | 6 +- .../service/nvdrv/devices/nvdisp_disp0.cpp | 6 +- .../hle/service/nvdrv/devices/nvdisp_disp0.h | 8 +-- .../service/nvdrv/devices/nvhost_as_gpu.cpp | 31 +++++----- .../hle/service/nvdrv/devices/nvhost_as_gpu.h | 30 +++++----- .../hle/service/nvdrv/devices/nvhost_ctrl.cpp | 19 +++--- .../hle/service/nvdrv/devices/nvhost_ctrl.h | 21 ++++--- .../service/nvdrv/devices/nvhost_ctrl_gpu.cpp | 32 +++++----- .../service/nvdrv/devices/nvhost_ctrl_gpu.h | 34 +++++------ .../hle/service/nvdrv/devices/nvhost_gpu.cpp | 59 ++++++++++--------- .../hle/service/nvdrv/devices/nvhost_gpu.h | 36 +++++------ .../service/nvdrv/devices/nvhost_nvdec.cpp | 6 +- .../hle/service/nvdrv/devices/nvhost_nvdec.h | 8 +-- .../nvdrv/devices/nvhost_nvdec_common.cpp | 15 +++-- .../nvdrv/devices/nvhost_nvdec_common.h | 12 ++-- .../service/nvdrv/devices/nvhost_nvjpg.cpp | 8 +-- .../hle/service/nvdrv/devices/nvhost_nvjpg.h | 10 ++-- .../hle/service/nvdrv/devices/nvhost_vic.cpp | 6 +- .../hle/service/nvdrv/devices/nvhost_vic.h | 8 +-- src/core/hle/service/nvdrv/devices/nvmap.cpp | 20 +++---- src/core/hle/service/nvdrv/devices/nvmap.h | 20 +++---- src/core/hle/service/nvdrv/nvdrv.cpp | 8 +-- src/core/hle/service/nvdrv/nvdrv.h | 8 +-- .../hle/service/nvdrv/nvdrv_interface.cpp | 24 ++++---- src/core/hle/service/nvdrv/nvdrv_interface.h | 3 + src/core/hle/service/nvnflinger/parcel.h | 7 ++- .../backend/glsl/glsl_emit_context.cpp | 2 +- .../backend/spirv/emit_spirv.cpp | 2 +- .../backend/spirv/spirv_emit_context.cpp | 2 +- src/shader_recompiler/runtime_info.h | 3 +- src/video_core/buffer_cache/buffer_cache.h | 4 +- .../buffer_cache/buffer_cache_base.h | 4 +- src/video_core/cdma_pusher.h | 1 - src/video_core/dma_pusher.h | 8 ++- src/video_core/engines/maxwell_dma.cpp | 35 ++++++----- src/video_core/host1x/codecs/h264.cpp | 4 +- src/video_core/memory_manager.cpp | 13 ++-- src/video_core/memory_manager.h | 15 +++-- .../renderer_opengl/gl_shader_cache.cpp | 4 +- .../renderer_vulkan/vk_buffer_cache.cpp | 2 +- .../renderer_vulkan/vk_pipeline_cache.cpp | 10 +++- .../renderer_vulkan/vk_texture_cache.cpp | 27 +++++---- src/video_core/shader_cache.cpp | 4 +- src/video_core/texture_cache/image_base.h | 5 +- src/video_core/texture_cache/texture_cache.h | 14 ++--- .../texture_cache/texture_cache_base.h | 4 +- src/video_core/texture_cache/util.cpp | 48 ++++++++------- src/video_core/texture_cache/util.h | 31 +++++----- src/video_core/transform_feedback.cpp | 8 ++- src/video_core/transform_feedback.h | 2 +- .../vulkan_common/vulkan_device.cpp | 1 + 84 files changed, 501 insertions(+), 458 deletions(-) diff --git a/src/audio_core/device/audio_buffers.h b/src/audio_core/device/audio_buffers.h index 15082f6c6..5d8ed0ef7 100644 --- a/src/audio_core/device/audio_buffers.h +++ b/src/audio_core/device/audio_buffers.h @@ -7,6 +7,7 @@ #include #include #include +#include #include "audio_buffer.h" #include "audio_core/device/device_session.h" @@ -48,7 +49,7 @@ public: * * @param out_buffers - The buffers which were registered. */ - void RegisterBuffers(std::vector& out_buffers) { + void RegisterBuffers(boost::container::static_vector& out_buffers) { std::scoped_lock l{lock}; const s32 to_register{std::min(std::min(appended_count, BufferAppendLimit), BufferAppendLimit - registered_count)}; @@ -162,7 +163,8 @@ public: * @param max_buffers - Maximum number of buffers to released. * @return The number of buffers released. */ - u32 GetRegisteredAppendedBuffers(std::vector& buffers_flushed, u32 max_buffers) { + u32 GetRegisteredAppendedBuffers( + boost::container::static_vector& buffers_flushed, u32 max_buffers) { std::scoped_lock l{lock}; if (registered_count + appended_count == 0) { return 0; @@ -270,7 +272,7 @@ public: */ bool FlushBuffers(u32& buffers_released) { std::scoped_lock l{lock}; - std::vector buffers_flushed{}; + boost::container::static_vector buffers_flushed{}; buffers_released = GetRegisteredAppendedBuffers(buffers_flushed, append_limit); diff --git a/src/audio_core/device/device_session.cpp b/src/audio_core/device/device_session.cpp index b5c0ef0e6..86811fcb8 100644 --- a/src/audio_core/device/device_session.cpp +++ b/src/audio_core/device/device_session.cpp @@ -79,7 +79,7 @@ void DeviceSession::ClearBuffers() { } } -void DeviceSession::AppendBuffers(std::span buffers) const { +void DeviceSession::AppendBuffers(std::span buffers) { for (const auto& buffer : buffers) { Sink::SinkBuffer new_buffer{ .frames = buffer.size / (channel_count * sizeof(s16)), @@ -88,13 +88,13 @@ void DeviceSession::AppendBuffers(std::span buffers) const { .consumed = false, }; + tmp_samples.resize_destructive(buffer.size / sizeof(s16)); if (type == Sink::StreamType::In) { - std::vector samples{}; - stream->AppendBuffer(new_buffer, samples); + stream->AppendBuffer(new_buffer, tmp_samples); } else { - std::vector samples(buffer.size / sizeof(s16)); - system.ApplicationMemory().ReadBlockUnsafe(buffer.samples, samples.data(), buffer.size); - stream->AppendBuffer(new_buffer, samples); + system.ApplicationMemory().ReadBlockUnsafe(buffer.samples, tmp_samples.data(), + buffer.size); + stream->AppendBuffer(new_buffer, tmp_samples); } } } diff --git a/src/audio_core/device/device_session.h b/src/audio_core/device/device_session.h index 75f766c68..7d52f362d 100644 --- a/src/audio_core/device/device_session.h +++ b/src/audio_core/device/device_session.h @@ -10,6 +10,7 @@ #include "audio_core/common/common.h" #include "audio_core/sink/sink.h" +#include "common/scratch_buffer.h" #include "core/hle/service/audio/errors.h" namespace Core { @@ -62,7 +63,7 @@ public: * * @param buffers - The buffers to play. */ - void AppendBuffers(std::span buffers) const; + void AppendBuffers(std::span buffers); /** * (Audio In only) Pop samples from the backend, and write them back to this buffer's address. @@ -146,8 +147,8 @@ private: std::shared_ptr thread_event; /// Is this session initialised? bool initialized{}; - /// Buffer queue - std::vector buffer_queue{}; + /// Temporary sample buffer + Common::ScratchBuffer tmp_samples{}; }; } // namespace AudioCore diff --git a/src/audio_core/in/audio_in_system.cpp b/src/audio_core/in/audio_in_system.cpp index e23e51758..579129121 100644 --- a/src/audio_core/in/audio_in_system.cpp +++ b/src/audio_core/in/audio_in_system.cpp @@ -2,6 +2,7 @@ // SPDX-License-Identifier: GPL-2.0-or-later #include + #include "audio_core/audio_event.h" #include "audio_core/audio_manager.h" #include "audio_core/in/audio_in_system.h" @@ -89,7 +90,7 @@ Result System::Start() { session->Start(); state = State::Started; - std::vector buffers_to_flush{}; + boost::container::static_vector buffers_to_flush{}; buffers.RegisterBuffers(buffers_to_flush); session->AppendBuffers(buffers_to_flush); session->SetRingSize(static_cast(buffers_to_flush.size())); @@ -134,7 +135,7 @@ bool System::AppendBuffer(const AudioInBuffer& buffer, const u64 tag) { void System::RegisterBuffers() { if (state == State::Started) { - std::vector registered_buffers{}; + boost::container::static_vector registered_buffers{}; buffers.RegisterBuffers(registered_buffers); session->AppendBuffers(registered_buffers); } diff --git a/src/audio_core/out/audio_out_system.cpp b/src/audio_core/out/audio_out_system.cpp index bd13f7219..0adf64bd3 100644 --- a/src/audio_core/out/audio_out_system.cpp +++ b/src/audio_core/out/audio_out_system.cpp @@ -89,7 +89,7 @@ Result System::Start() { session->Start(); state = State::Started; - std::vector buffers_to_flush{}; + boost::container::static_vector buffers_to_flush{}; buffers.RegisterBuffers(buffers_to_flush); session->AppendBuffers(buffers_to_flush); session->SetRingSize(static_cast(buffers_to_flush.size())); @@ -134,7 +134,7 @@ bool System::AppendBuffer(const AudioOutBuffer& buffer, u64 tag) { void System::RegisterBuffers() { if (state == State::Started) { - std::vector registered_buffers{}; + boost::container::static_vector registered_buffers{}; buffers.RegisterBuffers(registered_buffers); session->AppendBuffers(registered_buffers); } diff --git a/src/audio_core/renderer/command/data_source/decode.cpp b/src/audio_core/renderer/command/data_source/decode.cpp index ff5d31bd6..f45933203 100644 --- a/src/audio_core/renderer/command/data_source/decode.cpp +++ b/src/audio_core/renderer/command/data_source/decode.cpp @@ -8,6 +8,7 @@ #include "audio_core/renderer/command/resample/resample.h" #include "common/fixed_point.h" #include "common/logging/log.h" +#include "common/scratch_buffer.h" #include "core/memory.h" namespace AudioCore::AudioRenderer { @@ -27,6 +28,7 @@ constexpr std::array PitchBySrcQuality = {4, 8, 4}; template static u32 DecodePcm(Core::Memory::Memory& memory, std::span out_buffer, const DecodeArg& req) { + std::array tmp_samples{}; constexpr s32 min{std::numeric_limits::min()}; constexpr s32 max{std::numeric_limits::max()}; @@ -49,18 +51,17 @@ static u32 DecodePcm(Core::Memory::Memory& memory, std::span out_buffer, const u64 size{channel_count * samples_to_decode}; const u64 size_bytes{size * sizeof(T)}; - std::vector samples(size); - memory.ReadBlockUnsafe(source, samples.data(), size_bytes); + memory.ReadBlockUnsafe(source, tmp_samples.data(), size_bytes); if constexpr (std::is_floating_point_v) { for (u32 i = 0; i < samples_to_decode; i++) { - auto sample{static_cast(samples[i * channel_count + req.target_channel] * + auto sample{static_cast(tmp_samples[i * channel_count + req.target_channel] * std::numeric_limits::max())}; out_buffer[i] = static_cast(std::clamp(sample, min, max)); } } else { for (u32 i = 0; i < samples_to_decode; i++) { - out_buffer[i] = samples[i * channel_count + req.target_channel]; + out_buffer[i] = tmp_samples[i * channel_count + req.target_channel]; } } } break; @@ -73,17 +74,16 @@ static u32 DecodePcm(Core::Memory::Memory& memory, std::span out_buffer, } const VAddr source{req.buffer + ((req.start_offset + req.offset) * sizeof(T))}; - std::vector samples(samples_to_decode); - memory.ReadBlockUnsafe(source, samples.data(), samples_to_decode * sizeof(T)); + memory.ReadBlockUnsafe(source, tmp_samples.data(), samples_to_decode * sizeof(T)); if constexpr (std::is_floating_point_v) { for (u32 i = 0; i < samples_to_decode; i++) { - auto sample{static_cast(samples[i * channel_count + req.target_channel] * + auto sample{static_cast(tmp_samples[i * channel_count + req.target_channel] * std::numeric_limits::max())}; out_buffer[i] = static_cast(std::clamp(sample, min, max)); } } else { - std::memcpy(out_buffer.data(), samples.data(), samples_to_decode * sizeof(s16)); + std::memcpy(out_buffer.data(), tmp_samples.data(), samples_to_decode * sizeof(s16)); } break; } @@ -101,6 +101,7 @@ static u32 DecodePcm(Core::Memory::Memory& memory, std::span out_buffer, */ static u32 DecodeAdpcm(Core::Memory::Memory& memory, std::span out_buffer, const DecodeArg& req) { + std::array wavebuffer{}; constexpr u32 SamplesPerFrame{14}; constexpr u32 NibblesPerFrame{16}; @@ -138,9 +139,7 @@ static u32 DecodeAdpcm(Core::Memory::Memory& memory, std::span out_buffer, } const auto size{std::max((samples_to_process / 8U) * SamplesPerFrame, 8U)}; - std::vector wavebuffer(size); - memory.ReadBlockUnsafe(req.buffer + position_in_frame / 2, wavebuffer.data(), - wavebuffer.size()); + memory.ReadBlockUnsafe(req.buffer + position_in_frame / 2, wavebuffer.data(), size); auto context{req.adpcm_context}; auto header{context->header}; @@ -258,7 +257,7 @@ void DecodeFromWaveBuffers(Core::Memory::Memory& memory, const DecodeFromWaveBuf u32 offset{voice_state.offset}; auto output_buffer{args.output}; - std::vector temp_buffer(TempBufferSize, 0); + std::array temp_buffer{}; while (remaining_sample_count > 0) { const auto samples_to_write{std::min(remaining_sample_count, max_remaining_sample_count)}; diff --git a/src/audio_core/renderer/command/effect/compressor.cpp b/src/audio_core/renderer/command/effect/compressor.cpp index 7229618e8..ee9b68d5b 100644 --- a/src/audio_core/renderer/command/effect/compressor.cpp +++ b/src/audio_core/renderer/command/effect/compressor.cpp @@ -44,8 +44,8 @@ static void InitializeCompressorEffect(const CompressorInfo::ParameterVersion2& static void ApplyCompressorEffect(const CompressorInfo::ParameterVersion2& params, CompressorInfo::State& state, bool enabled, - std::vector> input_buffers, - std::vector> output_buffers, u32 sample_count) { + std::span> input_buffers, + std::span> output_buffers, u32 sample_count) { if (enabled) { auto state_00{state.unk_00}; auto state_04{state.unk_04}; @@ -124,8 +124,8 @@ void CompressorCommand::Dump([[maybe_unused]] const ADSP::CommandListProcessor& } void CompressorCommand::Process(const ADSP::CommandListProcessor& processor) { - std::vector> input_buffers(parameter.channel_count); - std::vector> output_buffers(parameter.channel_count); + std::array, MaxChannels> input_buffers{}; + std::array, MaxChannels> output_buffers{}; for (s16 i = 0; i < parameter.channel_count; i++) { input_buffers[i] = processor.mix_buffers.subspan(inputs[i] * processor.sample_count, diff --git a/src/audio_core/renderer/command/effect/delay.cpp b/src/audio_core/renderer/command/effect/delay.cpp index a4e408d40..e536cbb1e 100644 --- a/src/audio_core/renderer/command/effect/delay.cpp +++ b/src/audio_core/renderer/command/effect/delay.cpp @@ -51,7 +51,7 @@ static void InitializeDelayEffect(const DelayInfo::ParameterVersion1& params, state.delay_lines[channel].sample_count_max = sample_count_max.to_int_floor(); state.delay_lines[channel].sample_count = sample_count.to_int_floor(); state.delay_lines[channel].buffer.resize(state.delay_lines[channel].sample_count, 0); - if (state.delay_lines[channel].buffer.size() == 0) { + if (state.delay_lines[channel].sample_count == 0) { state.delay_lines[channel].buffer.push_back(0); } state.delay_lines[channel].buffer_pos = 0; @@ -74,8 +74,8 @@ static void InitializeDelayEffect(const DelayInfo::ParameterVersion1& params, */ template static void ApplyDelay(const DelayInfo::ParameterVersion1& params, DelayInfo::State& state, - std::vector>& inputs, - std::vector>& outputs, const u32 sample_count) { + std::span> inputs, std::span> outputs, + const u32 sample_count) { for (u32 sample_index = 0; sample_index < sample_count; sample_index++) { std::array, NumChannels> input_samples{}; for (u32 channel = 0; channel < NumChannels; channel++) { @@ -153,8 +153,8 @@ static void ApplyDelay(const DelayInfo::ParameterVersion1& params, DelayInfo::St * @param sample_count - Number of samples to process. */ static void ApplyDelayEffect(const DelayInfo::ParameterVersion1& params, DelayInfo::State& state, - const bool enabled, std::vector>& inputs, - std::vector>& outputs, const u32 sample_count) { + const bool enabled, std::span> inputs, + std::span> outputs, const u32 sample_count) { if (!IsChannelCountValid(params.channel_count)) { LOG_ERROR(Service_Audio, "Invalid delay channels {}", params.channel_count); @@ -208,8 +208,8 @@ void DelayCommand::Dump([[maybe_unused]] const ADSP::CommandListProcessor& proce } void DelayCommand::Process(const ADSP::CommandListProcessor& processor) { - std::vector> input_buffers(parameter.channel_count); - std::vector> output_buffers(parameter.channel_count); + std::array, MaxChannels> input_buffers{}; + std::array, MaxChannels> output_buffers{}; for (s16 i = 0; i < parameter.channel_count; i++) { input_buffers[i] = processor.mix_buffers.subspan(inputs[i] * processor.sample_count, diff --git a/src/audio_core/renderer/command/effect/i3dl2_reverb.cpp b/src/audio_core/renderer/command/effect/i3dl2_reverb.cpp index 27d8b9844..d2bfb67cc 100644 --- a/src/audio_core/renderer/command/effect/i3dl2_reverb.cpp +++ b/src/audio_core/renderer/command/effect/i3dl2_reverb.cpp @@ -408,8 +408,8 @@ void I3dl2ReverbCommand::Dump([[maybe_unused]] const ADSP::CommandListProcessor& } void I3dl2ReverbCommand::Process(const ADSP::CommandListProcessor& processor) { - std::vector> input_buffers(parameter.channel_count); - std::vector> output_buffers(parameter.channel_count); + std::array, MaxChannels> input_buffers{}; + std::array, MaxChannels> output_buffers{}; for (u32 i = 0; i < parameter.channel_count; i++) { input_buffers[i] = processor.mix_buffers.subspan(inputs[i] * processor.sample_count, diff --git a/src/audio_core/renderer/command/effect/light_limiter.cpp b/src/audio_core/renderer/command/effect/light_limiter.cpp index e8fb0e2fc..4161a9821 100644 --- a/src/audio_core/renderer/command/effect/light_limiter.cpp +++ b/src/audio_core/renderer/command/effect/light_limiter.cpp @@ -47,8 +47,8 @@ static void InitializeLightLimiterEffect(const LightLimiterInfo::ParameterVersio */ static void ApplyLightLimiterEffect(const LightLimiterInfo::ParameterVersion2& params, LightLimiterInfo::State& state, const bool enabled, - std::vector>& inputs, - std::vector>& outputs, const u32 sample_count, + std::span> inputs, + std::span> outputs, const u32 sample_count, LightLimiterInfo::StatisticsInternal* statistics) { constexpr s64 min{std::numeric_limits::min()}; constexpr s64 max{std::numeric_limits::max()}; @@ -147,8 +147,8 @@ void LightLimiterVersion1Command::Dump([[maybe_unused]] const ADSP::CommandListP } void LightLimiterVersion1Command::Process(const ADSP::CommandListProcessor& processor) { - std::vector> input_buffers(parameter.channel_count); - std::vector> output_buffers(parameter.channel_count); + std::array, MaxChannels> input_buffers{}; + std::array, MaxChannels> output_buffers{}; for (u32 i = 0; i < parameter.channel_count; i++) { input_buffers[i] = processor.mix_buffers.subspan(inputs[i] * processor.sample_count, @@ -190,8 +190,8 @@ void LightLimiterVersion2Command::Dump([[maybe_unused]] const ADSP::CommandListP } void LightLimiterVersion2Command::Process(const ADSP::CommandListProcessor& processor) { - std::vector> input_buffers(parameter.channel_count); - std::vector> output_buffers(parameter.channel_count); + std::array, MaxChannels> input_buffers{}; + std::array, MaxChannels> output_buffers{}; for (u32 i = 0; i < parameter.channel_count; i++) { input_buffers[i] = processor.mix_buffers.subspan(inputs[i] * processor.sample_count, diff --git a/src/audio_core/renderer/command/effect/reverb.cpp b/src/audio_core/renderer/command/effect/reverb.cpp index 8b9b65214..fc2f15a5e 100644 --- a/src/audio_core/renderer/command/effect/reverb.cpp +++ b/src/audio_core/renderer/command/effect/reverb.cpp @@ -250,8 +250,8 @@ static Common::FixedPoint<50, 14> Axfx2AllPassTick(ReverbInfo::ReverbDelayLine& */ template static void ApplyReverbEffect(const ReverbInfo::ParameterVersion2& params, ReverbInfo::State& state, - std::vector>& inputs, - std::vector>& outputs, const u32 sample_count) { + std::span> inputs, + std::span> outputs, const u32 sample_count) { static constexpr std::array OutTapIndexes1Ch{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }; @@ -369,8 +369,8 @@ static void ApplyReverbEffect(const ReverbInfo::ParameterVersion2& params, Rever * @param sample_count - Number of samples to process. */ static void ApplyReverbEffect(const ReverbInfo::ParameterVersion2& params, ReverbInfo::State& state, - const bool enabled, std::vector>& inputs, - std::vector>& outputs, const u32 sample_count) { + const bool enabled, std::span> inputs, + std::span> outputs, const u32 sample_count) { if (enabled) { switch (params.channel_count) { case 0: @@ -412,8 +412,8 @@ void ReverbCommand::Dump([[maybe_unused]] const ADSP::CommandListProcessor& proc } void ReverbCommand::Process(const ADSP::CommandListProcessor& processor) { - std::vector> input_buffers(parameter.channel_count); - std::vector> output_buffers(parameter.channel_count); + std::array, MaxChannels> input_buffers{}; + std::array, MaxChannels> output_buffers{}; for (u32 i = 0; i < parameter.channel_count; i++) { input_buffers[i] = processor.mix_buffers.subspan(inputs[i] * processor.sample_count, diff --git a/src/audio_core/renderer/command/sink/circular_buffer.cpp b/src/audio_core/renderer/command/sink/circular_buffer.cpp index ded5afc94..e2ce59792 100644 --- a/src/audio_core/renderer/command/sink/circular_buffer.cpp +++ b/src/audio_core/renderer/command/sink/circular_buffer.cpp @@ -24,7 +24,7 @@ void CircularBufferSinkCommand::Process(const ADSP::CommandListProcessor& proces constexpr s32 min{std::numeric_limits::min()}; constexpr s32 max{std::numeric_limits::max()}; - std::vector output(processor.sample_count); + std::array output{}; for (u32 channel = 0; channel < input_count; channel++) { auto input{processor.mix_buffers.subspan(inputs[channel] * processor.sample_count, processor.sample_count)}; @@ -33,7 +33,7 @@ void CircularBufferSinkCommand::Process(const ADSP::CommandListProcessor& proces } processor.memory->WriteBlockUnsafe(address + pos, output.data(), - output.size() * sizeof(s16)); + processor.sample_count * sizeof(s16)); pos += static_cast(processor.sample_count * sizeof(s16)); if (pos >= size) { pos = 0; diff --git a/src/audio_core/renderer/command/sink/device.cpp b/src/audio_core/renderer/command/sink/device.cpp index e88372a75..5f74dd7ad 100644 --- a/src/audio_core/renderer/command/sink/device.cpp +++ b/src/audio_core/renderer/command/sink/device.cpp @@ -33,8 +33,7 @@ void DeviceSinkCommand::Process(const ADSP::CommandListProcessor& processor) { .consumed{false}, }; - std::vector samples(out_buffer.frames * input_count); - + std::array samples{}; for (u32 channel = 0; channel < input_count; channel++) { const auto offset{inputs[channel] * out_buffer.frames}; @@ -45,7 +44,7 @@ void DeviceSinkCommand::Process(const ADSP::CommandListProcessor& processor) { } out_buffer.tag = reinterpret_cast(samples.data()); - stream->AppendBuffer(out_buffer, samples); + stream->AppendBuffer(out_buffer, {samples.data(), out_buffer.frames * input_count}); if (stream->IsPaused()) { stream->Start(); diff --git a/src/audio_core/renderer/mix/mix_context.cpp b/src/audio_core/renderer/mix/mix_context.cpp index 35b748ede..3a18ae7c2 100644 --- a/src/audio_core/renderer/mix/mix_context.cpp +++ b/src/audio_core/renderer/mix/mix_context.cpp @@ -125,10 +125,10 @@ bool MixContext::TSortInfo(const SplitterContext& splitter_context) { return false; } - std::vector sorted_results{node_states.GetSortedResuls()}; - const auto result_size{std::min(count, static_cast(sorted_results.size()))}; + auto sorted_results{node_states.GetSortedResuls()}; + const auto result_size{std::min(count, static_cast(sorted_results.second))}; for (s32 i = 0; i < result_size; i++) { - sorted_mix_infos[i] = &mix_infos[sorted_results[i]]; + sorted_mix_infos[i] = &mix_infos[sorted_results.first[i]]; } CalcMixBufferOffset(); diff --git a/src/audio_core/renderer/nodes/node_states.cpp b/src/audio_core/renderer/nodes/node_states.cpp index 1821a51e6..b7a44a54c 100644 --- a/src/audio_core/renderer/nodes/node_states.cpp +++ b/src/audio_core/renderer/nodes/node_states.cpp @@ -134,8 +134,8 @@ u32 NodeStates::GetNodeCount() const { return node_count; } -std::vector NodeStates::GetSortedResuls() const { - return {results.rbegin(), results.rbegin() + result_pos}; +std::pair::reverse_iterator, size_t> NodeStates::GetSortedResuls() const { + return {results.rbegin(), result_pos}; } } // namespace AudioCore::AudioRenderer diff --git a/src/audio_core/renderer/nodes/node_states.h b/src/audio_core/renderer/nodes/node_states.h index 94b1d1254..e768cd4b5 100644 --- a/src/audio_core/renderer/nodes/node_states.h +++ b/src/audio_core/renderer/nodes/node_states.h @@ -175,7 +175,7 @@ public: * * @return Vector of nodes in reverse order. */ - std::vector GetSortedResuls() const; + std::pair::reverse_iterator, size_t> GetSortedResuls() const; private: /// Number of nodes in the graph diff --git a/src/audio_core/renderer/system.cpp b/src/audio_core/renderer/system.cpp index 53b258c4f..a23627472 100644 --- a/src/audio_core/renderer/system.cpp +++ b/src/audio_core/renderer/system.cpp @@ -444,6 +444,7 @@ Result System::Update(std::span input, std::span performance, std: std::scoped_lock l{lock}; const auto start_time{core.CoreTiming().GetClockTicks()}; + std::memset(output.data(), 0, output.size()); InfoUpdater info_updater(input, output, process_handle, behavior); diff --git a/src/audio_core/sink/null_sink.h b/src/audio_core/sink/null_sink.h index 1215d3cd2..b6b43c93e 100644 --- a/src/audio_core/sink/null_sink.h +++ b/src/audio_core/sink/null_sink.h @@ -20,7 +20,7 @@ public: explicit NullSinkStreamImpl(Core::System& system_, StreamType type_) : SinkStream{system_, type_} {} ~NullSinkStreamImpl() override {} - void AppendBuffer(SinkBuffer&, std::vector&) override {} + void AppendBuffer(SinkBuffer&, std::span) override {} std::vector ReleaseBuffer(u64) override { return {}; } diff --git a/src/audio_core/sink/sink_stream.cpp b/src/audio_core/sink/sink_stream.cpp index 9a718a9cc..404dcd0e9 100644 --- a/src/audio_core/sink/sink_stream.cpp +++ b/src/audio_core/sink/sink_stream.cpp @@ -18,7 +18,7 @@ namespace AudioCore::Sink { -void SinkStream::AppendBuffer(SinkBuffer& buffer, std::vector& samples) { +void SinkStream::AppendBuffer(SinkBuffer& buffer, std::span samples) { if (type == StreamType::In) { queue.enqueue(buffer); queued_buffers++; @@ -66,15 +66,16 @@ void SinkStream::AppendBuffer(SinkBuffer& buffer, std::vector& samples) { static_cast(std::clamp(right_sample, min, max)); } - samples.resize(samples.size() / system_channels * device_channels); + samples = samples.subspan(0, samples.size() / system_channels * device_channels); } else if (system_channels == 2 && device_channels == 6) { // We need moar samples! Not all games will provide 6 channel audio. // TODO: Implement some upmixing here. Currently just passthrough, with other // channels left as silence. - std::vector new_samples(samples.size() / system_channels * device_channels, 0); + auto new_size = samples.size() / system_channels * device_channels; + tmp_samples.resize_destructive(new_size); - for (u32 read_index = 0, write_index = 0; read_index < samples.size(); + for (u32 read_index = 0, write_index = 0; read_index < new_size; read_index += system_channels, write_index += device_channels) { const auto left_sample{static_cast(std::clamp( static_cast( @@ -82,7 +83,7 @@ void SinkStream::AppendBuffer(SinkBuffer& buffer, std::vector& samples) { volume), min, max))}; - new_samples[write_index + static_cast(Channels::FrontLeft)] = left_sample; + tmp_samples[write_index + static_cast(Channels::FrontLeft)] = left_sample; const auto right_sample{static_cast(std::clamp( static_cast( @@ -90,9 +91,9 @@ void SinkStream::AppendBuffer(SinkBuffer& buffer, std::vector& samples) { volume), min, max))}; - new_samples[write_index + static_cast(Channels::FrontRight)] = right_sample; + tmp_samples[write_index + static_cast(Channels::FrontRight)] = right_sample; } - samples = std::move(new_samples); + samples = std::span(tmp_samples); } else if (volume != 1.0f) { for (u32 i = 0; i < samples.size(); i++) { diff --git a/src/audio_core/sink/sink_stream.h b/src/audio_core/sink/sink_stream.h index 41cbadc9c..98d72ace1 100644 --- a/src/audio_core/sink/sink_stream.h +++ b/src/audio_core/sink/sink_stream.h @@ -16,6 +16,7 @@ #include "common/polyfill_thread.h" #include "common/reader_writer_queue.h" #include "common/ring_buffer.h" +#include "common/scratch_buffer.h" #include "common/thread.h" namespace Core { @@ -170,7 +171,7 @@ public: * @param buffer - Audio buffer information to be queued. * @param samples - The s16 samples to be queue for playback. */ - virtual void AppendBuffer(SinkBuffer& buffer, std::vector& samples); + virtual void AppendBuffer(SinkBuffer& buffer, std::span samples); /** * Release a buffer. Audio In only, will fill a buffer with recorded samples. @@ -255,6 +256,8 @@ private: /// Signalled when ring buffer entries are consumed std::condition_variable_any release_cv; std::mutex release_mutex; + /// Temporary buffer for appending samples when upmixing + Common::ScratchBuffer tmp_samples{}; }; using SinkStreamPtr = std::unique_ptr; diff --git a/src/common/ring_buffer.h b/src/common/ring_buffer.h index 4c328ab44..416680d44 100644 --- a/src/common/ring_buffer.h +++ b/src/common/ring_buffer.h @@ -9,6 +9,7 @@ #include #include #include +#include #include #include @@ -53,7 +54,7 @@ public: return push_count; } - std::size_t Push(const std::vector& input) { + std::size_t Push(const std::span input) { return Push(input.data(), input.size()); } diff --git a/src/common/scratch_buffer.h b/src/common/scratch_buffer.h index a69a5a7af..6fe907953 100644 --- a/src/common/scratch_buffer.h +++ b/src/common/scratch_buffer.h @@ -3,6 +3,9 @@ #pragma once +#include + +#include "common/concepts.h" #include "common/make_unique_for_overwrite.h" namespace Common { @@ -16,6 +19,12 @@ namespace Common { template class ScratchBuffer { public: + using iterator = T*; + using const_iterator = const T*; + using value_type = T; + using element_type = T; + using iterator_category = std::contiguous_iterator_tag; + ScratchBuffer() = default; explicit ScratchBuffer(size_t initial_capacity) diff --git a/src/core/hle/kernel/k_synchronization_object.cpp b/src/core/hle/kernel/k_synchronization_object.cpp index b7da3eee7..3e5b735b1 100644 --- a/src/core/hle/kernel/k_synchronization_object.cpp +++ b/src/core/hle/kernel/k_synchronization_object.cpp @@ -3,6 +3,7 @@ #include "common/assert.h" #include "common/common_types.h" +#include "common/scratch_buffer.h" #include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h" #include "core/hle/kernel/k_synchronization_object.h" @@ -75,7 +76,7 @@ Result KSynchronizationObject::Wait(KernelCore& kernel, s32* out_index, KSynchronizationObject** objects, const s32 num_objects, s64 timeout) { // Allocate space on stack for thread nodes. - std::vector thread_nodes(num_objects); + std::array thread_nodes; // Prepare for wait. KThread* thread = GetCurrentThreadPointer(kernel); diff --git a/src/core/hle/kernel/k_thread.cpp b/src/core/hle/kernel/k_thread.cpp index 908811e2c..adb6ec581 100644 --- a/src/core/hle/kernel/k_thread.cpp +++ b/src/core/hle/kernel/k_thread.cpp @@ -909,7 +909,7 @@ Result KThread::SetActivity(Svc::ThreadActivity activity) { R_SUCCEED(); } -Result KThread::GetThreadContext3(std::vector& out) { +Result KThread::GetThreadContext3(Common::ScratchBuffer& out) { // Lock ourselves. KScopedLightLock lk{m_activity_pause_lock}; @@ -927,15 +927,13 @@ Result KThread::GetThreadContext3(std::vector& out) { // Mask away mode bits, interrupt bits, IL bit, and other reserved bits. auto context = GetContext64(); context.pstate &= 0xFF0FFE20; - - out.resize(sizeof(context)); + out.resize_destructive(sizeof(context)); std::memcpy(out.data(), std::addressof(context), sizeof(context)); } else { // Mask away mode bits, interrupt bits, IL bit, and other reserved bits. auto context = GetContext32(); context.cpsr &= 0xFF0FFE20; - - out.resize(sizeof(context)); + out.resize_destructive(sizeof(context)); std::memcpy(out.data(), std::addressof(context), sizeof(context)); } } diff --git a/src/core/hle/kernel/k_thread.h b/src/core/hle/kernel/k_thread.h index 37fe5db77..dd662b3f8 100644 --- a/src/core/hle/kernel/k_thread.h +++ b/src/core/hle/kernel/k_thread.h @@ -15,6 +15,7 @@ #include "common/intrusive_list.h" #include "common/intrusive_red_black_tree.h" +#include "common/scratch_buffer.h" #include "common/spin_lock.h" #include "core/arm/arm_interface.h" #include "core/hle/kernel/k_affinity_mask.h" @@ -567,7 +568,7 @@ public: void RemoveWaiter(KThread* thread); - Result GetThreadContext3(std::vector& out); + Result GetThreadContext3(Common::ScratchBuffer& out); KThread* RemoveUserWaiterByKey(bool* out_has_waiters, KProcessAddress key) { return this->RemoveWaiterByKey(out_has_waiters, key, false); diff --git a/src/core/hle/kernel/svc/svc_ipc.cpp b/src/core/hle/kernel/svc/svc_ipc.cpp index ea03068aa..60247df2e 100644 --- a/src/core/hle/kernel/svc/svc_ipc.cpp +++ b/src/core/hle/kernel/svc/svc_ipc.cpp @@ -2,6 +2,7 @@ // SPDX-License-Identifier: GPL-2.0-or-later #include "common/scope_exit.h" +#include "common/scratch_buffer.h" #include "core/core.h" #include "core/hle/kernel/k_client_session.h" #include "core/hle/kernel/k_process.h" @@ -45,11 +46,11 @@ Result ReplyAndReceive(Core::System& system, s32* out_index, uint64_t handles_ad handles_addr, static_cast(sizeof(Handle) * num_handles)), ResultInvalidPointer); - std::vector handles(num_handles); + std::array handles; GetCurrentMemory(kernel).ReadBlock(handles_addr, handles.data(), sizeof(Handle) * num_handles); // Convert handle list to object table. - std::vector objs(num_handles); + std::array objs; R_UNLESS(handle_table.GetMultipleObjects(objs.data(), handles.data(), num_handles), ResultInvalidHandle); @@ -80,7 +81,7 @@ Result ReplyAndReceive(Core::System& system, s32* out_index, uint64_t handles_ad // Wait for an object. s32 index; Result result = KSynchronizationObject::Wait(kernel, std::addressof(index), objs.data(), - static_cast(objs.size()), timeout_ns); + num_handles, timeout_ns); if (result == ResultTimedOut) { R_RETURN(result); } diff --git a/src/core/hle/kernel/svc/svc_synchronization.cpp b/src/core/hle/kernel/svc/svc_synchronization.cpp index 04d65f0bd..53df5bcd8 100644 --- a/src/core/hle/kernel/svc/svc_synchronization.cpp +++ b/src/core/hle/kernel/svc/svc_synchronization.cpp @@ -2,6 +2,7 @@ // SPDX-License-Identifier: GPL-2.0-or-later #include "common/scope_exit.h" +#include "common/scratch_buffer.h" #include "core/core.h" #include "core/hle/kernel/k_process.h" #include "core/hle/kernel/k_readable_event.h" @@ -54,7 +55,7 @@ static Result WaitSynchronization(Core::System& system, int32_t* out_index, cons // Get the synchronization context. auto& kernel = system.Kernel(); auto& handle_table = GetCurrentProcess(kernel).GetHandleTable(); - std::vector objs(num_handles); + std::array objs; // Copy user handles. if (num_handles > 0) { @@ -72,8 +73,8 @@ static Result WaitSynchronization(Core::System& system, int32_t* out_index, cons }); // Wait on the objects. - Result res = KSynchronizationObject::Wait(kernel, out_index, objs.data(), - static_cast(objs.size()), timeout_ns); + Result res = + KSynchronizationObject::Wait(kernel, out_index, objs.data(), num_handles, timeout_ns); R_SUCCEED_IF(res == ResultSessionClosed); R_RETURN(res); @@ -87,8 +88,7 @@ Result WaitSynchronization(Core::System& system, int32_t* out_index, u64 user_ha // Ensure number of handles is valid. R_UNLESS(0 <= num_handles && num_handles <= Svc::ArgumentHandleCountMax, ResultOutOfRange); - - std::vector handles(num_handles); + std::array handles; if (num_handles > 0) { GetCurrentMemory(system.Kernel()) .ReadBlock(user_handles, handles.data(), num_handles * sizeof(Handle)); diff --git a/src/core/hle/kernel/svc/svc_thread.cpp b/src/core/hle/kernel/svc/svc_thread.cpp index 37b54079c..36b94e6bf 100644 --- a/src/core/hle/kernel/svc/svc_thread.cpp +++ b/src/core/hle/kernel/svc/svc_thread.cpp @@ -174,7 +174,7 @@ Result GetThreadContext3(Core::System& system, u64 out_context, Handle thread_ha } // Get the thread context. - std::vector context; + static thread_local Common::ScratchBuffer context; R_TRY(thread->GetThreadContext3(context)); // Copy the thread context to user space. diff --git a/src/core/hle/service/audio/audin_u.cpp b/src/core/hle/service/audio/audin_u.cpp index f0640c64f..c8d574993 100644 --- a/src/core/hle/service/audio/audin_u.cpp +++ b/src/core/hle/service/audio/audin_u.cpp @@ -5,6 +5,7 @@ #include "audio_core/renderer/audio_device.h" #include "common/common_funcs.h" #include "common/logging/log.h" +#include "common/settings.h" #include "common/string_util.h" #include "core/core.h" #include "core/hle/kernel/k_event.h" @@ -123,19 +124,13 @@ private: void GetReleasedAudioInBuffer(HLERequestContext& ctx) { const auto write_buffer_size = ctx.GetWriteBufferNumElements(); - std::vector released_buffers(write_buffer_size); + tmp_buffer.resize_destructive(write_buffer_size); + tmp_buffer[0] = 0; - const auto count = impl->GetReleasedBuffers(released_buffers); + const auto count = impl->GetReleasedBuffers(tmp_buffer); - [[maybe_unused]] std::string tags{}; - for (u32 i = 0; i < count; i++) { - tags += fmt::format("{:08X}, ", released_buffers[i]); - } - [[maybe_unused]] auto sessionid{impl->GetSystem().GetSessionId()}; - LOG_TRACE(Service_Audio, "called. Session {} released {} buffers: {}", sessionid, count, - tags); + ctx.WriteBuffer(tmp_buffer); - ctx.WriteBuffer(released_buffers); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(ResultSuccess); rb.Push(count); @@ -200,6 +195,7 @@ private: KernelHelpers::ServiceContext service_context; Kernel::KEvent* event; std::shared_ptr impl; + Common::ScratchBuffer tmp_buffer; }; AudInU::AudInU(Core::System& system_) diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp index 3e62fa4fc..032c8c11f 100644 --- a/src/core/hle/service/audio/audout_u.cpp +++ b/src/core/hle/service/audio/audout_u.cpp @@ -123,19 +123,13 @@ private: void GetReleasedAudioOutBuffers(HLERequestContext& ctx) { const auto write_buffer_size = ctx.GetWriteBufferNumElements(); - std::vector released_buffers(write_buffer_size); + tmp_buffer.resize_destructive(write_buffer_size); + tmp_buffer[0] = 0; - const auto count = impl->GetReleasedBuffers(released_buffers); + const auto count = impl->GetReleasedBuffers(tmp_buffer); - [[maybe_unused]] std::string tags{}; - for (u32 i = 0; i < count; i++) { - tags += fmt::format("{:08X}, ", released_buffers[i]); - } - [[maybe_unused]] const auto sessionid{impl->GetSystem().GetSessionId()}; - LOG_TRACE(Service_Audio, "called. Session {} released {} buffers: {}", sessionid, count, - tags); + ctx.WriteBuffer(tmp_buffer); - ctx.WriteBuffer(released_buffers); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(ResultSuccess); rb.Push(count); @@ -211,6 +205,7 @@ private: KernelHelpers::ServiceContext service_context; Kernel::KEvent* event; std::shared_ptr impl; + Common::ScratchBuffer tmp_buffer; }; AudOutU::AudOutU(Core::System& system_) diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp index 7086d4750..12845c23a 100644 --- a/src/core/hle/service/audio/audren_u.cpp +++ b/src/core/hle/service/audio/audren_u.cpp @@ -116,28 +116,26 @@ private: // These buffers are written manually to avoid an issue with WriteBuffer throwing errors for // checking size 0. Performance size is 0 for most games. - std::vector output{}; - std::vector performance{}; auto is_buffer_b{ctx.BufferDescriptorB()[0].Size() != 0}; if (is_buffer_b) { const auto buffersB{ctx.BufferDescriptorB()}; - output.resize(buffersB[0].Size(), 0); - performance.resize(buffersB[1].Size(), 0); + tmp_output.resize_destructive(buffersB[0].Size()); + tmp_performance.resize_destructive(buffersB[1].Size()); } else { const auto buffersC{ctx.BufferDescriptorC()}; - output.resize(buffersC[0].Size(), 0); - performance.resize(buffersC[1].Size(), 0); + tmp_output.resize_destructive(buffersC[0].Size()); + tmp_performance.resize_destructive(buffersC[1].Size()); } - auto result = impl->RequestUpdate(input, performance, output); + auto result = impl->RequestUpdate(input, tmp_performance, tmp_output); if (result.IsSuccess()) { if (is_buffer_b) { - ctx.WriteBufferB(output.data(), output.size(), 0); - ctx.WriteBufferB(performance.data(), performance.size(), 1); + ctx.WriteBufferB(tmp_output.data(), tmp_output.size(), 0); + ctx.WriteBufferB(tmp_performance.data(), tmp_performance.size(), 1); } else { - ctx.WriteBufferC(output.data(), output.size(), 0); - ctx.WriteBufferC(performance.data(), performance.size(), 1); + ctx.WriteBufferC(tmp_output.data(), tmp_output.size(), 0); + ctx.WriteBufferC(tmp_performance.data(), tmp_performance.size(), 1); } } else { LOG_ERROR(Service_Audio, "RequestUpdate failed error 0x{:02X}!", result.description); @@ -235,6 +233,8 @@ private: Kernel::KEvent* rendered_event; Manager& manager; std::unique_ptr impl; + Common::ScratchBuffer tmp_output; + Common::ScratchBuffer tmp_performance; }; class IAudioDevice final : public ServiceFramework { diff --git a/src/core/hle/service/audio/audren_u.h b/src/core/hle/service/audio/audren_u.h index 24ce37e87..d8e9c8719 100644 --- a/src/core/hle/service/audio/audren_u.h +++ b/src/core/hle/service/audio/audren_u.h @@ -4,6 +4,7 @@ #pragma once #include "audio_core/audio_render_manager.h" +#include "common/scratch_buffer.h" #include "core/hle/service/kernel_helpers.h" #include "core/hle/service/service.h" diff --git a/src/core/hle/service/audio/hwopus.cpp b/src/core/hle/service/audio/hwopus.cpp index 451ac224a..c835f6cb7 100644 --- a/src/core/hle/service/audio/hwopus.cpp +++ b/src/core/hle/service/audio/hwopus.cpp @@ -68,13 +68,13 @@ private: ExtraBehavior extra_behavior) { u32 consumed = 0; u32 sample_count = 0; - std::vector samples(ctx.GetWriteBufferNumElements()); + tmp_samples.resize_destructive(ctx.GetWriteBufferNumElements()); if (extra_behavior == ExtraBehavior::ResetContext) { ResetDecoderContext(); } - if (!DecodeOpusData(consumed, sample_count, ctx.ReadBuffer(), samples, performance)) { + if (!DecodeOpusData(consumed, sample_count, ctx.ReadBuffer(), tmp_samples, performance)) { LOG_ERROR(Audio, "Failed to decode opus data"); IPC::ResponseBuilder rb{ctx, 2}; // TODO(ogniK): Use correct error code @@ -90,11 +90,11 @@ private: if (performance) { rb.Push(*performance); } - ctx.WriteBuffer(samples); + ctx.WriteBuffer(tmp_samples); } bool DecodeOpusData(u32& consumed, u32& sample_count, std::span input, - std::vector& output, u64* out_performance_time) const { + std::span output, u64* out_performance_time) const { const auto start_time = std::chrono::steady_clock::now(); const std::size_t raw_output_sz = output.size() * sizeof(opus_int16); if (sizeof(OpusPacketHeader) > input.size()) { @@ -154,6 +154,7 @@ private: OpusDecoderPtr decoder; u32 sample_rate; u32 channel_count; + Common::ScratchBuffer tmp_samples; }; class IHardwareOpusDecoderManager final : public ServiceFramework { diff --git a/src/core/hle/service/nvdrv/devices/nvdevice.h b/src/core/hle/service/nvdrv/devices/nvdevice.h index ab1f30f9e..a04538d5d 100644 --- a/src/core/hle/service/nvdrv/devices/nvdevice.h +++ b/src/core/hle/service/nvdrv/devices/nvdevice.h @@ -34,7 +34,7 @@ public: * @returns The result code of the ioctl. */ virtual NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) = 0; + std::span output) = 0; /** * Handles an ioctl2 request. @@ -45,7 +45,7 @@ public: * @returns The result code of the ioctl. */ virtual NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) = 0; + std::span inline_input, std::span output) = 0; /** * Handles an ioctl3 request. @@ -56,7 +56,7 @@ public: * @returns The result code of the ioctl. */ virtual NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) = 0; + std::span output, std::span inline_output) = 0; /** * Called once a device is opened diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp index 0fe242e9d..05a43d8dc 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp @@ -18,19 +18,19 @@ nvdisp_disp0::nvdisp_disp0(Core::System& system_, NvCore::Container& core) nvdisp_disp0::~nvdisp_disp0() = default; NvResult nvdisp_disp0::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } NvResult nvdisp_disp0::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } NvResult nvdisp_disp0::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { + std::span output, std::span inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h index bcd0e3ed5..daee05fe8 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h @@ -26,11 +26,11 @@ public: ~nvdisp_disp0() override; NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp index 681bd0867..07e570a9f 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp @@ -28,7 +28,7 @@ nvhost_as_gpu::nvhost_as_gpu(Core::System& system_, Module& module_, NvCore::Con nvhost_as_gpu::~nvhost_as_gpu() = default; NvResult nvhost_as_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { switch (command.group) { case 'A': switch (command.cmd) { @@ -61,13 +61,13 @@ NvResult nvhost_as_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span i } NvResult nvhost_as_gpu::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } NvResult nvhost_as_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { + std::span output, std::span inline_output) { switch (command.group) { case 'A': switch (command.cmd) { @@ -87,7 +87,7 @@ NvResult nvhost_as_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span i void nvhost_as_gpu::OnOpen(DeviceFD fd) {} void nvhost_as_gpu::OnClose(DeviceFD fd) {} -NvResult nvhost_as_gpu::AllocAsEx(std::span input, std::vector& output) { +NvResult nvhost_as_gpu::AllocAsEx(std::span input, std::span output) { IoctlAllocAsEx params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -141,7 +141,7 @@ NvResult nvhost_as_gpu::AllocAsEx(std::span input, std::vector& ou return NvResult::Success; } -NvResult nvhost_as_gpu::AllocateSpace(std::span input, std::vector& output) { +NvResult nvhost_as_gpu::AllocateSpace(std::span input, std::span output) { IoctlAllocSpace params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -220,7 +220,7 @@ void nvhost_as_gpu::FreeMappingLocked(u64 offset) { mapping_map.erase(offset); } -NvResult nvhost_as_gpu::FreeSpace(std::span input, std::vector& output) { +NvResult nvhost_as_gpu::FreeSpace(std::span input, std::span output) { IoctlFreeSpace params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -266,15 +266,14 @@ NvResult nvhost_as_gpu::FreeSpace(std::span input, std::vector& ou return NvResult::Success; } -NvResult nvhost_as_gpu::Remap(std::span input, std::vector& output) { +NvResult nvhost_as_gpu::Remap(std::span input, std::span output) { const auto num_entries = input.size() / sizeof(IoctlRemapEntry); LOG_DEBUG(Service_NVDRV, "called, num_entries=0x{:X}", num_entries); - std::vector entries(num_entries); - std::memcpy(entries.data(), input.data(), input.size()); - std::scoped_lock lock(mutex); + entries.resize_destructive(num_entries); + std::memcpy(entries.data(), input.data(), input.size()); if (!vm.initialised) { return NvResult::BadValue; @@ -320,7 +319,7 @@ NvResult nvhost_as_gpu::Remap(std::span input, std::vector& output return NvResult::Success; } -NvResult nvhost_as_gpu::MapBufferEx(std::span input, std::vector& output) { +NvResult nvhost_as_gpu::MapBufferEx(std::span input, std::span output) { IoctlMapBufferEx params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -424,7 +423,7 @@ NvResult nvhost_as_gpu::MapBufferEx(std::span input, std::vector& return NvResult::Success; } -NvResult nvhost_as_gpu::UnmapBuffer(std::span input, std::vector& output) { +NvResult nvhost_as_gpu::UnmapBuffer(std::span input, std::span output) { IoctlUnmapBuffer params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -463,7 +462,7 @@ NvResult nvhost_as_gpu::UnmapBuffer(std::span input, std::vector& return NvResult::Success; } -NvResult nvhost_as_gpu::BindChannel(std::span input, std::vector& output) { +NvResult nvhost_as_gpu::BindChannel(std::span input, std::span output) { IoctlBindChannel params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, fd={:X}", params.fd); @@ -492,7 +491,7 @@ void nvhost_as_gpu::GetVARegionsImpl(IoctlGetVaRegions& params) { }; } -NvResult nvhost_as_gpu::GetVARegions(std::span input, std::vector& output) { +NvResult nvhost_as_gpu::GetVARegions(std::span input, std::span output) { IoctlGetVaRegions params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -511,8 +510,8 @@ NvResult nvhost_as_gpu::GetVARegions(std::span input, std::vector& return NvResult::Success; } -NvResult nvhost_as_gpu::GetVARegions(std::span input, std::vector& output, - std::vector& inline_output) { +NvResult nvhost_as_gpu::GetVARegions(std::span input, std::span output, + std::span inline_output) { IoctlGetVaRegions params{}; std::memcpy(¶ms, input.data(), input.size()); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h index 1aba8d579..2af3e1260 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h @@ -15,6 +15,7 @@ #include "common/address_space.h" #include "common/common_funcs.h" #include "common/common_types.h" +#include "common/scratch_buffer.h" #include "common/swap.h" #include "core/hle/service/nvdrv/core/nvmap.h" #include "core/hle/service/nvdrv/devices/nvdevice.h" @@ -48,11 +49,11 @@ public: ~nvhost_as_gpu() override; NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; @@ -138,18 +139,18 @@ private: static_assert(sizeof(IoctlGetVaRegions) == 16 + sizeof(VaRegion) * 2, "IoctlGetVaRegions is incorrect size"); - NvResult AllocAsEx(std::span input, std::vector& output); - NvResult AllocateSpace(std::span input, std::vector& output); - NvResult Remap(std::span input, std::vector& output); - NvResult MapBufferEx(std::span input, std::vector& output); - NvResult UnmapBuffer(std::span input, std::vector& output); - NvResult FreeSpace(std::span input, std::vector& output); - NvResult BindChannel(std::span input, std::vector& output); + NvResult AllocAsEx(std::span input, std::span output); + NvResult AllocateSpace(std::span input, std::span output); + NvResult Remap(std::span input, std::span output); + NvResult MapBufferEx(std::span input, std::span output); + NvResult UnmapBuffer(std::span input, std::span output); + NvResult FreeSpace(std::span input, std::span output); + NvResult BindChannel(std::span input, std::span output); void GetVARegionsImpl(IoctlGetVaRegions& params); - NvResult GetVARegions(std::span input, std::vector& output); - NvResult GetVARegions(std::span input, std::vector& output, - std::vector& inline_output); + NvResult GetVARegions(std::span input, std::span output); + NvResult GetVARegions(std::span input, std::span output, + std::span inline_output); void FreeMappingLocked(u64 offset); @@ -212,6 +213,7 @@ private: bool initialised{}; } vm; std::shared_ptr gmmu; + Common::ScratchBuffer entries; // s32 channel{}; // u32 big_page_size{VM::DEFAULT_BIG_PAGE_SIZE}; diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp index e12025560..4d55554b4 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp @@ -35,7 +35,7 @@ nvhost_ctrl::~nvhost_ctrl() { } NvResult nvhost_ctrl::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { switch (command.group) { case 0x0: switch (command.cmd) { @@ -64,13 +64,13 @@ NvResult nvhost_ctrl::Ioctl1(DeviceFD fd, Ioctl command, std::span inp } NvResult nvhost_ctrl::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } NvResult nvhost_ctrl::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_outpu) { + std::span output, std::span inline_outpu) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } @@ -79,7 +79,7 @@ void nvhost_ctrl::OnOpen(DeviceFD fd) {} void nvhost_ctrl::OnClose(DeviceFD fd) {} -NvResult nvhost_ctrl::NvOsGetConfigU32(std::span input, std::vector& output) { +NvResult nvhost_ctrl::NvOsGetConfigU32(std::span input, std::span output) { IocGetConfigParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_TRACE(Service_NVDRV, "called, setting={}!{}", params.domain_str.data(), @@ -87,7 +87,7 @@ NvResult nvhost_ctrl::NvOsGetConfigU32(std::span input, std::vector input, std::vector& output, +NvResult nvhost_ctrl::IocCtrlEventWait(std::span input, std::span output, bool is_allocation) { IocCtrlEventWaitParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); @@ -231,7 +231,7 @@ NvResult nvhost_ctrl::FreeEvent(u32 slot) { return NvResult::Success; } -NvResult nvhost_ctrl::IocCtrlEventRegister(std::span input, std::vector& output) { +NvResult nvhost_ctrl::IocCtrlEventRegister(std::span input, std::span output) { IocCtrlEventRegisterParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); const u32 event_id = params.user_event_id; @@ -252,7 +252,7 @@ NvResult nvhost_ctrl::IocCtrlEventRegister(std::span input, std::vecto return NvResult::Success; } -NvResult nvhost_ctrl::IocCtrlEventUnregister(std::span input, std::vector& output) { +NvResult nvhost_ctrl::IocCtrlEventUnregister(std::span input, std::span output) { IocCtrlEventUnregisterParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); const u32 event_id = params.user_event_id & 0x00FF; @@ -262,8 +262,7 @@ NvResult nvhost_ctrl::IocCtrlEventUnregister(std::span input, std::vec return FreeEvent(event_id); } -NvResult nvhost_ctrl::IocCtrlEventUnregisterBatch(std::span input, - std::vector& output) { +NvResult nvhost_ctrl::IocCtrlEventUnregisterBatch(std::span input, std::span output) { IocCtrlEventUnregisterBatchParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); u64 event_mask = params.user_events; @@ -281,7 +280,7 @@ NvResult nvhost_ctrl::IocCtrlEventUnregisterBatch(std::span input, return NvResult::Success; } -NvResult nvhost_ctrl::IocCtrlClearEventWait(std::span input, std::vector& output) { +NvResult nvhost_ctrl::IocCtrlClearEventWait(std::span input, std::span output) { IocCtrlEventClearParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h index dd2e7888a..2efed4862 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h @@ -26,11 +26,11 @@ public: ~nvhost_ctrl() override; NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; @@ -186,13 +186,12 @@ private: static_assert(sizeof(IocCtrlEventUnregisterBatchParams) == 8, "IocCtrlEventKill is incorrect size"); - NvResult NvOsGetConfigU32(std::span input, std::vector& output); - NvResult IocCtrlEventWait(std::span input, std::vector& output, - bool is_allocation); - NvResult IocCtrlEventRegister(std::span input, std::vector& output); - NvResult IocCtrlEventUnregister(std::span input, std::vector& output); - NvResult IocCtrlEventUnregisterBatch(std::span input, std::vector& output); - NvResult IocCtrlClearEventWait(std::span input, std::vector& output); + NvResult NvOsGetConfigU32(std::span input, std::span output); + NvResult IocCtrlEventWait(std::span input, std::span output, bool is_allocation); + NvResult IocCtrlEventRegister(std::span input, std::span output); + NvResult IocCtrlEventUnregister(std::span input, std::span output); + NvResult IocCtrlEventUnregisterBatch(std::span input, std::span output); + NvResult IocCtrlClearEventWait(std::span input, std::span output); NvResult FreeEvent(u32 slot); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp index be3c083db..6081d92e9 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp @@ -22,7 +22,7 @@ nvhost_ctrl_gpu::~nvhost_ctrl_gpu() { } NvResult nvhost_ctrl_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { switch (command.group) { case 'G': switch (command.cmd) { @@ -54,13 +54,13 @@ NvResult nvhost_ctrl_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span } NvResult nvhost_ctrl_gpu::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } NvResult nvhost_ctrl_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { + std::span output, std::span inline_output) { switch (command.group) { case 'G': switch (command.cmd) { @@ -82,7 +82,7 @@ NvResult nvhost_ctrl_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span void nvhost_ctrl_gpu::OnOpen(DeviceFD fd) {} void nvhost_ctrl_gpu::OnClose(DeviceFD fd) {} -NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span input, std::vector& output) { +NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span input, std::span output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlCharacteristics params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -127,8 +127,8 @@ NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span input, std::vec return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span input, std::vector& output, - std::vector& inline_output) { +NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span input, std::span output, + std::span inline_output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlCharacteristics params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -175,7 +175,7 @@ NvResult nvhost_ctrl_gpu::GetCharacteristics(std::span input, std::vec return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span input, std::vector& output) { +NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span input, std::span output) { IoctlGpuGetTpcMasksArgs params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size); @@ -186,8 +186,8 @@ NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span input, std::vector return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span input, std::vector& output, - std::vector& inline_output) { +NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span input, std::span output, + std::span inline_output) { IoctlGpuGetTpcMasksArgs params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size); @@ -199,7 +199,7 @@ NvResult nvhost_ctrl_gpu::GetTPCMasks(std::span input, std::vector return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetActiveSlotMask(std::span input, std::vector& output) { +NvResult nvhost_ctrl_gpu::GetActiveSlotMask(std::span input, std::span output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlActiveSlotMask params{}; @@ -212,7 +212,7 @@ NvResult nvhost_ctrl_gpu::GetActiveSlotMask(std::span input, std::vect return NvResult::Success; } -NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(std::span input, std::vector& output) { +NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(std::span input, std::span output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlZcullGetCtxSize params{}; @@ -224,7 +224,7 @@ NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(std::span input, std::vector return NvResult::Success; } -NvResult nvhost_ctrl_gpu::ZCullGetInfo(std::span input, std::vector& output) { +NvResult nvhost_ctrl_gpu::ZCullGetInfo(std::span input, std::span output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlNvgpuGpuZcullGetInfoArgs params{}; @@ -247,7 +247,7 @@ NvResult nvhost_ctrl_gpu::ZCullGetInfo(std::span input, std::vector input, std::vector& output) { +NvResult nvhost_ctrl_gpu::ZBCSetTable(std::span input, std::span output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlZbcSetTable params{}; @@ -263,7 +263,7 @@ NvResult nvhost_ctrl_gpu::ZBCSetTable(std::span input, std::vector return NvResult::Success; } -NvResult nvhost_ctrl_gpu::ZBCQueryTable(std::span input, std::vector& output) { +NvResult nvhost_ctrl_gpu::ZBCQueryTable(std::span input, std::span output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlZbcQueryTable params{}; @@ -273,7 +273,7 @@ NvResult nvhost_ctrl_gpu::ZBCQueryTable(std::span input, std::vector input, std::vector& output) { +NvResult nvhost_ctrl_gpu::FlushL2(std::span input, std::span output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlFlushL2 params{}; @@ -283,7 +283,7 @@ NvResult nvhost_ctrl_gpu::FlushL2(std::span input, std::vector& ou return NvResult::Success; } -NvResult nvhost_ctrl_gpu::GetGpuTime(std::span input, std::vector& output) { +NvResult nvhost_ctrl_gpu::GetGpuTime(std::span input, std::span output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlGetGpuTime params{}; diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h index b9333d9d3..97995551c 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h @@ -22,11 +22,11 @@ public: ~nvhost_ctrl_gpu() override; NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; @@ -151,21 +151,21 @@ private: }; static_assert(sizeof(IoctlGetGpuTime) == 0x10, "IoctlGetGpuTime is incorrect size"); - NvResult GetCharacteristics(std::span input, std::vector& output); - NvResult GetCharacteristics(std::span input, std::vector& output, - std::vector& inline_output); + NvResult GetCharacteristics(std::span input, std::span output); + NvResult GetCharacteristics(std::span input, std::span output, + std::span inline_output); - NvResult GetTPCMasks(std::span input, std::vector& output); - NvResult GetTPCMasks(std::span input, std::vector& output, - std::vector& inline_output); + NvResult GetTPCMasks(std::span input, std::span output); + NvResult GetTPCMasks(std::span input, std::span output, + std::span inline_output); - NvResult GetActiveSlotMask(std::span input, std::vector& output); - NvResult ZCullGetCtxSize(std::span input, std::vector& output); - NvResult ZCullGetInfo(std::span input, std::vector& output); - NvResult ZBCSetTable(std::span input, std::vector& output); - NvResult ZBCQueryTable(std::span input, std::vector& output); - NvResult FlushL2(std::span input, std::vector& output); - NvResult GetGpuTime(std::span input, std::vector& output); + NvResult GetActiveSlotMask(std::span input, std::span output); + NvResult ZCullGetCtxSize(std::span input, std::span output); + NvResult ZCullGetInfo(std::span input, std::span output); + NvResult ZBCSetTable(std::span input, std::span output); + NvResult ZBCQueryTable(std::span input, std::span output); + NvResult FlushL2(std::span input, std::span output); + NvResult GetGpuTime(std::span input, std::span output); EventInterface& events_interface; diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp index 453a965dc..46a25fcab 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp @@ -47,7 +47,7 @@ nvhost_gpu::~nvhost_gpu() { } NvResult nvhost_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { switch (command.group) { case 0x0: switch (command.cmd) { @@ -99,7 +99,7 @@ NvResult nvhost_gpu::Ioctl1(DeviceFD fd, Ioctl command, std::span inpu }; NvResult nvhost_gpu::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { switch (command.group) { case 'H': switch (command.cmd) { @@ -113,7 +113,7 @@ NvResult nvhost_gpu::Ioctl2(DeviceFD fd, Ioctl command, std::span inpu } NvResult nvhost_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { + std::span output, std::span inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } @@ -121,7 +121,7 @@ NvResult nvhost_gpu::Ioctl3(DeviceFD fd, Ioctl command, std::span inpu void nvhost_gpu::OnOpen(DeviceFD fd) {} void nvhost_gpu::OnClose(DeviceFD fd) {} -NvResult nvhost_gpu::SetNVMAPfd(std::span input, std::vector& output) { +NvResult nvhost_gpu::SetNVMAPfd(std::span input, std::span output) { IoctlSetNvmapFD params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); @@ -130,7 +130,7 @@ NvResult nvhost_gpu::SetNVMAPfd(std::span input, std::vector& outp return NvResult::Success; } -NvResult nvhost_gpu::SetClientData(std::span input, std::vector& output) { +NvResult nvhost_gpu::SetClientData(std::span input, std::span output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlClientData params{}; @@ -139,7 +139,7 @@ NvResult nvhost_gpu::SetClientData(std::span input, std::vector& o return NvResult::Success; } -NvResult nvhost_gpu::GetClientData(std::span input, std::vector& output) { +NvResult nvhost_gpu::GetClientData(std::span input, std::span output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlClientData params{}; @@ -149,7 +149,7 @@ NvResult nvhost_gpu::GetClientData(std::span input, std::vector& o return NvResult::Success; } -NvResult nvhost_gpu::ZCullBind(std::span input, std::vector& output) { +NvResult nvhost_gpu::ZCullBind(std::span input, std::span output) { std::memcpy(&zcull_params, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, gpu_va={:X}, mode={:X}", zcull_params.gpu_va, zcull_params.mode); @@ -158,7 +158,7 @@ NvResult nvhost_gpu::ZCullBind(std::span input, std::vector& outpu return NvResult::Success; } -NvResult nvhost_gpu::SetErrorNotifier(std::span input, std::vector& output) { +NvResult nvhost_gpu::SetErrorNotifier(std::span input, std::span output) { IoctlSetErrorNotifier params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, "(STUBBED) called, offset={:X}, size={:X}, mem={:X}", params.offset, @@ -168,14 +168,14 @@ NvResult nvhost_gpu::SetErrorNotifier(std::span input, std::vector return NvResult::Success; } -NvResult nvhost_gpu::SetChannelPriority(std::span input, std::vector& output) { +NvResult nvhost_gpu::SetChannelPriority(std::span input, std::span output) { std::memcpy(&channel_priority, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "(STUBBED) called, priority={:X}", channel_priority); return NvResult::Success; } -NvResult nvhost_gpu::AllocGPFIFOEx2(std::span input, std::vector& output) { +NvResult nvhost_gpu::AllocGPFIFOEx2(std::span input, std::span output) { IoctlAllocGpfifoEx2 params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, @@ -197,7 +197,7 @@ NvResult nvhost_gpu::AllocGPFIFOEx2(std::span input, std::vector& return NvResult::Success; } -NvResult nvhost_gpu::AllocateObjectContext(std::span input, std::vector& output) { +NvResult nvhost_gpu::AllocateObjectContext(std::span input, std::span output) { IoctlAllocObjCtx params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, "(STUBBED) called, class_num={:X}, flags={:X}", params.class_num, @@ -208,7 +208,8 @@ NvResult nvhost_gpu::AllocateObjectContext(std::span input, std::vecto return NvResult::Success; } -static std::vector BuildWaitCommandList(NvFence fence) { +static boost::container::small_vector BuildWaitCommandList( + NvFence fence) { return { Tegra::BuildCommandHeader(Tegra::BufferMethods::SyncpointPayload, 1, Tegra::SubmissionMode::Increasing), @@ -219,35 +220,35 @@ static std::vector BuildWaitCommandList(NvFence fence) { }; } -static std::vector BuildIncrementCommandList(NvFence fence) { - std::vector result{ +static boost::container::small_vector BuildIncrementCommandList( + NvFence fence) { + boost::container::small_vector result{ Tegra::BuildCommandHeader(Tegra::BufferMethods::SyncpointPayload, 1, Tegra::SubmissionMode::Increasing), {}}; for (u32 count = 0; count < 2; ++count) { - result.emplace_back(Tegra::BuildCommandHeader(Tegra::BufferMethods::SyncpointOperation, 1, - Tegra::SubmissionMode::Increasing)); - result.emplace_back( + result.push_back(Tegra::BuildCommandHeader(Tegra::BufferMethods::SyncpointOperation, 1, + Tegra::SubmissionMode::Increasing)); + result.push_back( BuildFenceAction(Tegra::Engines::Puller::FenceOperation::Increment, fence.id)); } return result; } -static std::vector BuildIncrementWithWfiCommandList(NvFence fence) { - std::vector result{ +static boost::container::small_vector BuildIncrementWithWfiCommandList( + NvFence fence) { + boost::container::small_vector result{ Tegra::BuildCommandHeader(Tegra::BufferMethods::WaitForIdle, 1, Tegra::SubmissionMode::Increasing), {}}; - const std::vector increment{BuildIncrementCommandList(fence)}; - - result.insert(result.end(), increment.begin(), increment.end()); - + auto increment_list{BuildIncrementCommandList(fence)}; + result.insert(result.end(), increment_list.begin(), increment_list.end()); return result; } -NvResult nvhost_gpu::SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector& output, +NvResult nvhost_gpu::SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::span output, Tegra::CommandList&& entries) { LOG_TRACE(Service_NVDRV, "called, gpfifo={:X}, num_entries={:X}, flags={:X}", params.address, params.num_entries, params.flags.raw); @@ -293,7 +294,7 @@ NvResult nvhost_gpu::SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector return NvResult::Success; } -NvResult nvhost_gpu::SubmitGPFIFOBase(std::span input, std::vector& output, +NvResult nvhost_gpu::SubmitGPFIFOBase(std::span input, std::span output, bool kickoff) { if (input.size() < sizeof(IoctlSubmitGpfifo)) { UNIMPLEMENTED(); @@ -315,7 +316,7 @@ NvResult nvhost_gpu::SubmitGPFIFOBase(std::span input, std::vector } NvResult nvhost_gpu::SubmitGPFIFOBase(std::span input, std::span input_inline, - std::vector& output) { + std::span output) { if (input.size() < sizeof(IoctlSubmitGpfifo)) { UNIMPLEMENTED(); return NvResult::InvalidSize; @@ -327,7 +328,7 @@ NvResult nvhost_gpu::SubmitGPFIFOBase(std::span input, std::span input, std::vector& output) { +NvResult nvhost_gpu::GetWaitbase(std::span input, std::span output) { IoctlGetWaitbase params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlGetWaitbase)); LOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown); @@ -337,7 +338,7 @@ NvResult nvhost_gpu::GetWaitbase(std::span input, std::vector& out return NvResult::Success; } -NvResult nvhost_gpu::ChannelSetTimeout(std::span input, std::vector& output) { +NvResult nvhost_gpu::ChannelSetTimeout(std::span input, std::span output) { IoctlChannelSetTimeout params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlChannelSetTimeout)); LOG_INFO(Service_NVDRV, "called, timeout=0x{:X}", params.timeout); @@ -345,7 +346,7 @@ NvResult nvhost_gpu::ChannelSetTimeout(std::span input, std::vector input, std::vector& output) { +NvResult nvhost_gpu::ChannelSetTimeslice(std::span input, std::span output) { IoctlSetTimeslice params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlSetTimeslice)); LOG_INFO(Service_NVDRV, "called, timeslice=0x{:X}", params.timeslice); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h index 3ca58202d..529c20526 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h @@ -41,11 +41,11 @@ public: ~nvhost_gpu() override; NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; @@ -186,23 +186,23 @@ private: u32_le channel_priority{}; u32_le channel_timeslice{}; - NvResult SetNVMAPfd(std::span input, std::vector& output); - NvResult SetClientData(std::span input, std::vector& output); - NvResult GetClientData(std::span input, std::vector& output); - NvResult ZCullBind(std::span input, std::vector& output); - NvResult SetErrorNotifier(std::span input, std::vector& output); - NvResult SetChannelPriority(std::span input, std::vector& output); - NvResult AllocGPFIFOEx2(std::span input, std::vector& output); - NvResult AllocateObjectContext(std::span input, std::vector& output); - NvResult SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector& output, + NvResult SetNVMAPfd(std::span input, std::span output); + NvResult SetClientData(std::span input, std::span output); + NvResult GetClientData(std::span input, std::span output); + NvResult ZCullBind(std::span input, std::span output); + NvResult SetErrorNotifier(std::span input, std::span output); + NvResult SetChannelPriority(std::span input, std::span output); + NvResult AllocGPFIFOEx2(std::span input, std::span output); + NvResult AllocateObjectContext(std::span input, std::span output); + NvResult SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::span output, Tegra::CommandList&& entries); - NvResult SubmitGPFIFOBase(std::span input, std::vector& output, + NvResult SubmitGPFIFOBase(std::span input, std::span output, bool kickoff = false); NvResult SubmitGPFIFOBase(std::span input, std::span input_inline, - std::vector& output); - NvResult GetWaitbase(std::span input, std::vector& output); - NvResult ChannelSetTimeout(std::span input, std::vector& output); - NvResult ChannelSetTimeslice(std::span input, std::vector& output); + std::span output); + NvResult GetWaitbase(std::span input, std::span output); + NvResult ChannelSetTimeout(std::span input, std::span output); + NvResult ChannelSetTimeslice(std::span input, std::span output); EventInterface& events_interface; NvCore::Container& core; diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp index dc45169ad..a174442a6 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp @@ -16,7 +16,7 @@ nvhost_nvdec::nvhost_nvdec(Core::System& system_, NvCore::Container& core_) nvhost_nvdec::~nvhost_nvdec() = default; NvResult nvhost_nvdec::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { switch (command.group) { case 0x0: switch (command.cmd) { @@ -56,13 +56,13 @@ NvResult nvhost_nvdec::Ioctl1(DeviceFD fd, Ioctl command, std::span in } NvResult nvhost_nvdec::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } NvResult nvhost_nvdec::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { + std::span output, std::span inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h index 0d615bbcb..ad2233c49 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h @@ -14,11 +14,11 @@ public: ~nvhost_nvdec() override; NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp index 1ab51f10b..61649aa4a 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp @@ -36,7 +36,7 @@ std::size_t SliceVectors(std::span input, std::vector& dst, std::si // Writes the data in src to an offset into the dst vector. The offset is specified in bytes // Returns the number of bytes written into dst. template -std::size_t WriteVectors(std::vector& dst, const std::vector& src, std::size_t offset) { +std::size_t WriteVectors(std::span dst, const std::vector& src, std::size_t offset) { if (src.empty()) { return 0; } @@ -72,8 +72,7 @@ NvResult nvhost_nvdec_common::SetNVMAPfd(std::span input) { return NvResult::Success; } -NvResult nvhost_nvdec_common::Submit(DeviceFD fd, std::span input, - std::vector& output) { +NvResult nvhost_nvdec_common::Submit(DeviceFD fd, std::span input, std::span output) { IoctlSubmit params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlSubmit)); LOG_DEBUG(Service_NVDRV, "called NVDEC Submit, cmd_buffer_count={}", params.cmd_buffer_count); @@ -121,7 +120,7 @@ NvResult nvhost_nvdec_common::Submit(DeviceFD fd, std::span input, return NvResult::Success; } -NvResult nvhost_nvdec_common::GetSyncpoint(std::span input, std::vector& output) { +NvResult nvhost_nvdec_common::GetSyncpoint(std::span input, std::span output) { IoctlGetSyncpoint params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlGetSyncpoint)); LOG_DEBUG(Service_NVDRV, "called GetSyncpoint, id={}", params.param); @@ -133,7 +132,7 @@ NvResult nvhost_nvdec_common::GetSyncpoint(std::span input, std::vecto return NvResult::Success; } -NvResult nvhost_nvdec_common::GetWaitbase(std::span input, std::vector& output) { +NvResult nvhost_nvdec_common::GetWaitbase(std::span input, std::span output) { IoctlGetWaitbase params{}; LOG_CRITICAL(Service_NVDRV, "called WAITBASE"); std::memcpy(¶ms, input.data(), sizeof(IoctlGetWaitbase)); @@ -142,7 +141,7 @@ NvResult nvhost_nvdec_common::GetWaitbase(std::span input, std::vector return NvResult::Success; } -NvResult nvhost_nvdec_common::MapBuffer(std::span input, std::vector& output) { +NvResult nvhost_nvdec_common::MapBuffer(std::span input, std::span output) { IoctlMapBuffer params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlMapBuffer)); std::vector cmd_buffer_handles(params.num_entries); @@ -159,7 +158,7 @@ NvResult nvhost_nvdec_common::MapBuffer(std::span input, std::vector input, std::vector& output) { +NvResult nvhost_nvdec_common::UnmapBuffer(std::span input, std::span output) { IoctlMapBuffer params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlMapBuffer)); std::vector cmd_buffer_handles(params.num_entries); @@ -173,7 +172,7 @@ NvResult nvhost_nvdec_common::UnmapBuffer(std::span input, std::vector return NvResult::Success; } -NvResult nvhost_nvdec_common::SetSubmitTimeout(std::span input, std::vector& output) { +NvResult nvhost_nvdec_common::SetSubmitTimeout(std::span input, std::span output) { std::memcpy(&submit_timeout, input.data(), input.size()); LOG_WARNING(Service_NVDRV, "(STUBBED) called"); return NvResult::Success; diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h index 5af26a26f..9bb573bfe 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h @@ -108,12 +108,12 @@ protected: /// Ioctl command implementations NvResult SetNVMAPfd(std::span input); - NvResult Submit(DeviceFD fd, std::span input, std::vector& output); - NvResult GetSyncpoint(std::span input, std::vector& output); - NvResult GetWaitbase(std::span input, std::vector& output); - NvResult MapBuffer(std::span input, std::vector& output); - NvResult UnmapBuffer(std::span input, std::vector& output); - NvResult SetSubmitTimeout(std::span input, std::vector& output); + NvResult Submit(DeviceFD fd, std::span input, std::span output); + NvResult GetSyncpoint(std::span input, std::span output); + NvResult GetWaitbase(std::span input, std::span output); + NvResult MapBuffer(std::span input, std::span output); + NvResult UnmapBuffer(std::span input, std::span output); + NvResult SetSubmitTimeout(std::span input, std::span output); Kernel::KEvent* QueryEvent(u32 event_id) override; diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp index 39f30e7c8..a05c8cdae 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp @@ -13,7 +13,7 @@ nvhost_nvjpg::nvhost_nvjpg(Core::System& system_) : nvdevice{system_} {} nvhost_nvjpg::~nvhost_nvjpg() = default; NvResult nvhost_nvjpg::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { switch (command.group) { case 'H': switch (command.cmd) { @@ -32,13 +32,13 @@ NvResult nvhost_nvjpg::Ioctl1(DeviceFD fd, Ioctl command, std::span in } NvResult nvhost_nvjpg::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } NvResult nvhost_nvjpg::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { + std::span output, std::span inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } @@ -46,7 +46,7 @@ NvResult nvhost_nvjpg::Ioctl3(DeviceFD fd, Ioctl command, std::span in void nvhost_nvjpg::OnOpen(DeviceFD fd) {} void nvhost_nvjpg::OnClose(DeviceFD fd) {} -NvResult nvhost_nvjpg::SetNVMAPfd(std::span input, std::vector& output) { +NvResult nvhost_nvjpg::SetNVMAPfd(std::span input, std::span output) { IoctlSetNvmapFD params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h index 41b57e872..5623e0d47 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h @@ -16,11 +16,11 @@ public: ~nvhost_nvjpg() override; NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; @@ -33,7 +33,7 @@ private: s32_le nvmap_fd{}; - NvResult SetNVMAPfd(std::span input, std::vector& output); + NvResult SetNVMAPfd(std::span input, std::span output); }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp index b0ea402a7..c0b8684c3 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp @@ -16,7 +16,7 @@ nvhost_vic::nvhost_vic(Core::System& system_, NvCore::Container& core_) nvhost_vic::~nvhost_vic() = default; NvResult nvhost_vic::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { switch (command.group) { case 0x0: switch (command.cmd) { @@ -56,13 +56,13 @@ NvResult nvhost_vic::Ioctl1(DeviceFD fd, Ioctl command, std::span inpu } NvResult nvhost_vic::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } NvResult nvhost_vic::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { + std::span output, std::span inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.h b/src/core/hle/service/nvdrv/devices/nvhost_vic.h index b5e350a83..cadbcb0a5 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_vic.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.h @@ -13,11 +13,11 @@ public: ~nvhost_vic(); NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; diff --git a/src/core/hle/service/nvdrv/devices/nvmap.cpp b/src/core/hle/service/nvdrv/devices/nvmap.cpp index 07417f045..e7f7e273b 100644 --- a/src/core/hle/service/nvdrv/devices/nvmap.cpp +++ b/src/core/hle/service/nvdrv/devices/nvmap.cpp @@ -26,7 +26,7 @@ nvmap::nvmap(Core::System& system_, NvCore::Container& container_) nvmap::~nvmap() = default; NvResult nvmap::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { switch (command.group) { case 0x1: switch (command.cmd) { @@ -55,13 +55,13 @@ NvResult nvmap::Ioctl1(DeviceFD fd, Ioctl command, std::span input, } NvResult nvmap::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } -NvResult nvmap::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { +NvResult nvmap::Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) { UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); return NvResult::NotImplemented; } @@ -69,7 +69,7 @@ NvResult nvmap::Ioctl3(DeviceFD fd, Ioctl command, std::span input, void nvmap::OnOpen(DeviceFD fd) {} void nvmap::OnClose(DeviceFD fd) {} -NvResult nvmap::IocCreate(std::span input, std::vector& output) { +NvResult nvmap::IocCreate(std::span input, std::span output) { IocCreateParams params; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_DEBUG(Service_NVDRV, "called, size=0x{:08X}", params.size); @@ -89,7 +89,7 @@ NvResult nvmap::IocCreate(std::span input, std::vector& output) { return NvResult::Success; } -NvResult nvmap::IocAlloc(std::span input, std::vector& output) { +NvResult nvmap::IocAlloc(std::span input, std::span output) { IocAllocParams params; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_DEBUG(Service_NVDRV, "called, addr={:X}", params.address); @@ -137,7 +137,7 @@ NvResult nvmap::IocAlloc(std::span input, std::vector& output) { return result; } -NvResult nvmap::IocGetId(std::span input, std::vector& output) { +NvResult nvmap::IocGetId(std::span input, std::span output) { IocGetIdParams params; std::memcpy(¶ms, input.data(), sizeof(params)); @@ -161,7 +161,7 @@ NvResult nvmap::IocGetId(std::span input, std::vector& output) { return NvResult::Success; } -NvResult nvmap::IocFromId(std::span input, std::vector& output) { +NvResult nvmap::IocFromId(std::span input, std::span output) { IocFromIdParams params; std::memcpy(¶ms, input.data(), sizeof(params)); @@ -192,7 +192,7 @@ NvResult nvmap::IocFromId(std::span input, std::vector& output) { return NvResult::Success; } -NvResult nvmap::IocParam(std::span input, std::vector& output) { +NvResult nvmap::IocParam(std::span input, std::span output) { enum class ParamTypes { Size = 1, Alignment = 2, Base = 3, Heap = 4, Kind = 5, Compr = 6 }; IocParamParams params; @@ -241,7 +241,7 @@ NvResult nvmap::IocParam(std::span input, std::vector& output) { return NvResult::Success; } -NvResult nvmap::IocFree(std::span input, std::vector& output) { +NvResult nvmap::IocFree(std::span input, std::span output) { IocFreeParams params; std::memcpy(¶ms, input.data(), sizeof(params)); diff --git a/src/core/hle/service/nvdrv/devices/nvmap.h b/src/core/hle/service/nvdrv/devices/nvmap.h index 82bd3b118..40c65b430 100644 --- a/src/core/hle/service/nvdrv/devices/nvmap.h +++ b/src/core/hle/service/nvdrv/devices/nvmap.h @@ -27,11 +27,11 @@ public: nvmap& operator=(const nvmap&) = delete; NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) override; + std::span output) override; NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) override; - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output) override; + std::span inline_input, std::span output) override; + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) override; void OnOpen(DeviceFD fd) override; void OnClose(DeviceFD fd) override; @@ -106,12 +106,12 @@ private: }; static_assert(sizeof(IocGetIdParams) == 8, "IocGetIdParams has wrong size"); - NvResult IocCreate(std::span input, std::vector& output); - NvResult IocAlloc(std::span input, std::vector& output); - NvResult IocGetId(std::span input, std::vector& output); - NvResult IocFromId(std::span input, std::vector& output); - NvResult IocParam(std::span input, std::vector& output); - NvResult IocFree(std::span input, std::vector& output); + NvResult IocCreate(std::span input, std::span output); + NvResult IocAlloc(std::span input, std::span output); + NvResult IocGetId(std::span input, std::span output); + NvResult IocFromId(std::span input, std::span output); + NvResult IocParam(std::span input, std::span output); + NvResult IocFree(std::span input, std::span output); NvCore::Container& container; NvCore::NvMap& file; diff --git a/src/core/hle/service/nvdrv/nvdrv.cpp b/src/core/hle/service/nvdrv/nvdrv.cpp index 3d774eec4..9e46ee8dd 100644 --- a/src/core/hle/service/nvdrv/nvdrv.cpp +++ b/src/core/hle/service/nvdrv/nvdrv.cpp @@ -130,7 +130,7 @@ DeviceFD Module::Open(const std::string& device_name) { } NvResult Module::Ioctl1(DeviceFD fd, Ioctl command, std::span input, - std::vector& output) { + std::span output) { if (fd < 0) { LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); return NvResult::InvalidState; @@ -147,7 +147,7 @@ NvResult Module::Ioctl1(DeviceFD fd, Ioctl command, std::span input, } NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output) { + std::span inline_input, std::span output) { if (fd < 0) { LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); return NvResult::InvalidState; @@ -163,8 +163,8 @@ NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, std::span input, return itr->second->Ioctl2(fd, command, input, inline_input, output); } -NvResult Module::Ioctl3(DeviceFD fd, Ioctl command, std::span input, - std::vector& output, std::vector& inline_output) { +NvResult Module::Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output) { if (fd < 0) { LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); return NvResult::InvalidState; diff --git a/src/core/hle/service/nvdrv/nvdrv.h b/src/core/hle/service/nvdrv/nvdrv.h index 668be742b..d8622b3ca 100644 --- a/src/core/hle/service/nvdrv/nvdrv.h +++ b/src/core/hle/service/nvdrv/nvdrv.h @@ -80,13 +80,13 @@ public: DeviceFD Open(const std::string& device_name); /// Sends an ioctl command to the specified file descriptor. - NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, std::vector& output); + NvResult Ioctl1(DeviceFD fd, Ioctl command, std::span input, std::span output); NvResult Ioctl2(DeviceFD fd, Ioctl command, std::span input, - std::span inline_input, std::vector& output); + std::span inline_input, std::span output); - NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::vector& output, - std::vector& inline_output); + NvResult Ioctl3(DeviceFD fd, Ioctl command, std::span input, std::span output, + std::span inline_output); /// Closes a device file descriptor and returns operation success. NvResult Close(DeviceFD fd); diff --git a/src/core/hle/service/nvdrv/nvdrv_interface.cpp b/src/core/hle/service/nvdrv/nvdrv_interface.cpp index d010a1e03..348207e25 100644 --- a/src/core/hle/service/nvdrv/nvdrv_interface.cpp +++ b/src/core/hle/service/nvdrv/nvdrv_interface.cpp @@ -63,12 +63,12 @@ void NVDRV::Ioctl1(HLERequestContext& ctx) { } // Check device - std::vector output_buffer(ctx.GetWriteBufferSize(0)); + tmp_output.resize_destructive(ctx.GetWriteBufferSize(0)); const auto input_buffer = ctx.ReadBuffer(0); - const auto nv_result = nvdrv->Ioctl1(fd, command, input_buffer, output_buffer); + const auto nv_result = nvdrv->Ioctl1(fd, command, input_buffer, tmp_output); if (command.is_out != 0) { - ctx.WriteBuffer(output_buffer); + ctx.WriteBuffer(tmp_output); } IPC::ResponseBuilder rb{ctx, 3}; @@ -90,12 +90,12 @@ void NVDRV::Ioctl2(HLERequestContext& ctx) { const auto input_buffer = ctx.ReadBuffer(0); const auto input_inlined_buffer = ctx.ReadBuffer(1); - std::vector output_buffer(ctx.GetWriteBufferSize(0)); + tmp_output.resize_destructive(ctx.GetWriteBufferSize(0)); const auto nv_result = - nvdrv->Ioctl2(fd, command, input_buffer, input_inlined_buffer, output_buffer); + nvdrv->Ioctl2(fd, command, input_buffer, input_inlined_buffer, tmp_output); if (command.is_out != 0) { - ctx.WriteBuffer(output_buffer); + ctx.WriteBuffer(tmp_output); } IPC::ResponseBuilder rb{ctx, 3}; @@ -116,14 +116,12 @@ void NVDRV::Ioctl3(HLERequestContext& ctx) { } const auto input_buffer = ctx.ReadBuffer(0); - std::vector output_buffer(ctx.GetWriteBufferSize(0)); - std::vector output_buffer_inline(ctx.GetWriteBufferSize(1)); - - const auto nv_result = - nvdrv->Ioctl3(fd, command, input_buffer, output_buffer, output_buffer_inline); + tmp_output.resize_destructive(ctx.GetWriteBufferSize(0)); + tmp_output_inline.resize_destructive(ctx.GetWriteBufferSize(1)); + const auto nv_result = nvdrv->Ioctl3(fd, command, input_buffer, tmp_output, tmp_output_inline); if (command.is_out != 0) { - ctx.WriteBuffer(output_buffer, 0); - ctx.WriteBuffer(output_buffer_inline, 1); + ctx.WriteBuffer(tmp_output, 0); + ctx.WriteBuffer(tmp_output_inline, 1); } IPC::ResponseBuilder rb{ctx, 3}; diff --git a/src/core/hle/service/nvdrv/nvdrv_interface.h b/src/core/hle/service/nvdrv/nvdrv_interface.h index 881ea1a6b..4b593ff90 100644 --- a/src/core/hle/service/nvdrv/nvdrv_interface.h +++ b/src/core/hle/service/nvdrv/nvdrv_interface.h @@ -4,6 +4,7 @@ #pragma once #include +#include "common/scratch_buffer.h" #include "core/hle/service/nvdrv/nvdrv.h" #include "core/hle/service/service.h" @@ -33,6 +34,8 @@ private: u64 pid{}; bool is_initialized{}; + Common::ScratchBuffer tmp_output; + Common::ScratchBuffer tmp_output_inline; }; } // namespace Service::Nvidia diff --git a/src/core/hle/service/nvnflinger/parcel.h b/src/core/hle/service/nvnflinger/parcel.h index fb56d75d7..23ba315a0 100644 --- a/src/core/hle/service/nvnflinger/parcel.h +++ b/src/core/hle/service/nvnflinger/parcel.h @@ -6,6 +6,7 @@ #include #include #include +#include #include "common/alignment.h" #include "common/assert.h" @@ -167,7 +168,7 @@ public: private: template requires(std::is_trivially_copyable_v) - void WriteImpl(const T& val, std::vector& buffer) { + void WriteImpl(const T& val, boost::container::small_vector& buffer) { const size_t aligned_size = Common::AlignUp(sizeof(T), 4); const size_t old_size = buffer.size(); buffer.resize(old_size + aligned_size); @@ -176,8 +177,8 @@ private: } private: - std::vector m_data_buffer; - std::vector m_object_buffer; + boost::container::small_vector m_data_buffer; + boost::container::small_vector m_object_buffer; }; } // namespace Service::android diff --git a/src/shader_recompiler/backend/glsl/glsl_emit_context.cpp b/src/shader_recompiler/backend/glsl/glsl_emit_context.cpp index c3c2281bb..9ff4028c2 100644 --- a/src/shader_recompiler/backend/glsl/glsl_emit_context.cpp +++ b/src/shader_recompiler/backend/glsl/glsl_emit_context.cpp @@ -479,7 +479,7 @@ void EmitContext::DefineGenericOutput(size_t index, u32 invocations) { const u32 remainder{4 - element}; const TransformFeedbackVarying* xfb_varying{}; const size_t xfb_varying_index{base_index + element}; - if (xfb_varying_index < runtime_info.xfb_varyings.size()) { + if (xfb_varying_index < runtime_info.xfb_count) { xfb_varying = &runtime_info.xfb_varyings[xfb_varying_index]; xfb_varying = xfb_varying->components > 0 ? xfb_varying : nullptr; } diff --git a/src/shader_recompiler/backend/spirv/emit_spirv.cpp b/src/shader_recompiler/backend/spirv/emit_spirv.cpp index 0f86a8004..34592a01f 100644 --- a/src/shader_recompiler/backend/spirv/emit_spirv.cpp +++ b/src/shader_recompiler/backend/spirv/emit_spirv.cpp @@ -387,7 +387,7 @@ void SetupSignedNanCapabilities(const Profile& profile, const IR::Program& progr } void SetupTransformFeedbackCapabilities(EmitContext& ctx, Id main_func) { - if (ctx.runtime_info.xfb_varyings.empty()) { + if (ctx.runtime_info.xfb_count == 0) { return; } ctx.AddCapability(spv::Capability::TransformFeedback); diff --git a/src/shader_recompiler/backend/spirv/spirv_emit_context.cpp b/src/shader_recompiler/backend/spirv/spirv_emit_context.cpp index fd15f47ea..bec5db173 100644 --- a/src/shader_recompiler/backend/spirv/spirv_emit_context.cpp +++ b/src/shader_recompiler/backend/spirv/spirv_emit_context.cpp @@ -160,7 +160,7 @@ void DefineGenericOutput(EmitContext& ctx, size_t index, std::optional invo const u32 remainder{4 - element}; const TransformFeedbackVarying* xfb_varying{}; const size_t xfb_varying_index{base_attr_index + element}; - if (xfb_varying_index < ctx.runtime_info.xfb_varyings.size()) { + if (xfb_varying_index < ctx.runtime_info.xfb_count) { xfb_varying = &ctx.runtime_info.xfb_varyings[xfb_varying_index]; xfb_varying = xfb_varying->components > 0 ? xfb_varying : nullptr; } diff --git a/src/shader_recompiler/runtime_info.h b/src/shader_recompiler/runtime_info.h index 3b63c249f..619c0b138 100644 --- a/src/shader_recompiler/runtime_info.h +++ b/src/shader_recompiler/runtime_info.h @@ -84,7 +84,8 @@ struct RuntimeInfo { bool glasm_use_storage_buffers{}; /// Transform feedback state for each varying - std::vector xfb_varyings; + std::array xfb_varyings{}; + u32 xfb_count{0}; }; } // namespace Shader diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 45977d578..58a45ab67 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -207,7 +207,7 @@ bool BufferCache

::DMACopy(GPUVAddr src_address, GPUVAddr dest_address, u64 am if (has_new_downloads) { memory_tracker.MarkRegionAsGpuModified(*cpu_dest_address, amount); } - tmp_buffer.resize(amount); + tmp_buffer.resize_destructive(amount); cpu_memory.ReadBlockUnsafe(*cpu_src_address, tmp_buffer.data(), amount); cpu_memory.WriteBlockUnsafe(*cpu_dest_address, tmp_buffer.data(), amount); return true; @@ -1279,7 +1279,7 @@ template typename BufferCache

::OverlapResult BufferCache

::ResolveOverlaps(VAddr cpu_addr, u32 wanted_size) { static constexpr int STREAM_LEAP_THRESHOLD = 16; - std::vector overlap_ids; + boost::container::small_vector overlap_ids; VAddr begin = cpu_addr; VAddr end = cpu_addr + wanted_size; int stream_score = 0; diff --git a/src/video_core/buffer_cache/buffer_cache_base.h b/src/video_core/buffer_cache/buffer_cache_base.h index 63a120f7a..fe6068cfe 100644 --- a/src/video_core/buffer_cache/buffer_cache_base.h +++ b/src/video_core/buffer_cache/buffer_cache_base.h @@ -229,7 +229,7 @@ class BufferCache : public VideoCommon::ChannelSetupCaches; struct OverlapResult { - std::vector ids; + boost::container::small_vector ids; VAddr begin; VAddr end; bool has_stream_leap = false; @@ -582,7 +582,7 @@ private: BufferId inline_buffer_id; std::array> CACHING_PAGEBITS)> page_table; - std::vector tmp_buffer; + Common::ScratchBuffer tmp_buffer; }; } // namespace VideoCommon diff --git a/src/video_core/cdma_pusher.h b/src/video_core/cdma_pusher.h index 83112dfce..7d660af47 100644 --- a/src/video_core/cdma_pusher.h +++ b/src/video_core/cdma_pusher.h @@ -63,7 +63,6 @@ struct ChCommand { }; using ChCommandHeaderList = std::vector; -using ChCommandList = std::vector; struct ThiRegisters { u32_le increment_syncpt{}; diff --git a/src/video_core/dma_pusher.h b/src/video_core/dma_pusher.h index 1cdb690ed..8a2784cdc 100644 --- a/src/video_core/dma_pusher.h +++ b/src/video_core/dma_pusher.h @@ -6,6 +6,7 @@ #include #include #include +#include #include #include "common/bit_field.h" @@ -102,11 +103,12 @@ 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_) + explicit CommandList( + boost::container::small_vector&& prefetch_command_list_) : prefetch_command_list{std::move(prefetch_command_list_)} {} - std::vector command_lists; - std::vector prefetch_command_list; + boost::container::small_vector command_lists; + boost::container::small_vector prefetch_command_list; }; /** diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp index ebe5536de..bc1eb41e7 100644 --- a/src/video_core/engines/maxwell_dma.cpp +++ b/src/video_core/engines/maxwell_dma.cpp @@ -108,9 +108,11 @@ void MaxwellDMA::Launch() { if (regs.launch_dma.remap_enable != 0 && is_const_a_dst) { ASSERT(regs.remap_const.component_size_minus_one == 3); accelerate.BufferClear(regs.offset_out, regs.line_length_in, regs.remap_consta_value); - std::vector tmp_buffer(regs.line_length_in, regs.remap_consta_value); + read_buffer.resize_destructive(regs.line_length_in * sizeof(u32)); + std::span span(reinterpret_cast(read_buffer.data()), regs.line_length_in); + std::ranges::fill(span, regs.remap_consta_value); memory_manager.WriteBlockUnsafe(regs.offset_out, - reinterpret_cast(tmp_buffer.data()), + reinterpret_cast(read_buffer.data()), regs.line_length_in * sizeof(u32)); } else { memory_manager.FlushCaching(); @@ -126,32 +128,32 @@ void MaxwellDMA::Launch() { UNIMPLEMENTED_IF(regs.line_length_in % 16 != 0); UNIMPLEMENTED_IF(regs.offset_in % 16 != 0); UNIMPLEMENTED_IF(regs.offset_out % 16 != 0); - std::vector tmp_buffer(16); + read_buffer.resize_destructive(16); for (u32 offset = 0; offset < regs.line_length_in; offset += 16) { memory_manager.ReadBlockUnsafe( convert_linear_2_blocklinear_addr(regs.offset_in + offset), - tmp_buffer.data(), tmp_buffer.size()); - memory_manager.WriteBlockCached(regs.offset_out + offset, tmp_buffer.data(), - tmp_buffer.size()); + read_buffer.data(), read_buffer.size()); + memory_manager.WriteBlockCached(regs.offset_out + offset, read_buffer.data(), + read_buffer.size()); } } else if (is_src_pitch && !is_dst_pitch) { UNIMPLEMENTED_IF(regs.line_length_in % 16 != 0); UNIMPLEMENTED_IF(regs.offset_in % 16 != 0); UNIMPLEMENTED_IF(regs.offset_out % 16 != 0); - std::vector tmp_buffer(16); + read_buffer.resize_destructive(16); for (u32 offset = 0; offset < regs.line_length_in; offset += 16) { - memory_manager.ReadBlockUnsafe(regs.offset_in + offset, tmp_buffer.data(), - tmp_buffer.size()); + memory_manager.ReadBlockUnsafe(regs.offset_in + offset, read_buffer.data(), + read_buffer.size()); memory_manager.WriteBlockCached( convert_linear_2_blocklinear_addr(regs.offset_out + offset), - tmp_buffer.data(), tmp_buffer.size()); + read_buffer.data(), read_buffer.size()); } } else { if (!accelerate.BufferCopy(regs.offset_in, regs.offset_out, regs.line_length_in)) { - std::vector tmp_buffer(regs.line_length_in); - memory_manager.ReadBlockUnsafe(regs.offset_in, tmp_buffer.data(), + read_buffer.resize_destructive(regs.line_length_in); + memory_manager.ReadBlockUnsafe(regs.offset_in, read_buffer.data(), regs.line_length_in); - memory_manager.WriteBlockCached(regs.offset_out, tmp_buffer.data(), + memory_manager.WriteBlockCached(regs.offset_out, read_buffer.data(), regs.line_length_in); } } @@ -171,7 +173,8 @@ void MaxwellDMA::CopyBlockLinearToPitch() { src_operand.address = regs.offset_in; DMA::BufferOperand dst_operand; - dst_operand.pitch = regs.pitch_out; + u32 abs_pitch_out = std::abs(static_cast(regs.pitch_out)); + dst_operand.pitch = abs_pitch_out; dst_operand.width = regs.line_length_in; dst_operand.height = regs.line_count; dst_operand.address = regs.offset_out; @@ -218,7 +221,7 @@ void MaxwellDMA::CopyBlockLinearToPitch() { const size_t src_size = CalculateSize(true, bytes_per_pixel, width, height, depth, block_height, block_depth); - const size_t dst_size = static_cast(regs.pitch_out) * regs.line_count; + const size_t dst_size = static_cast(abs_pitch_out) * regs.line_count; read_buffer.resize_destructive(src_size); write_buffer.resize_destructive(dst_size); @@ -227,7 +230,7 @@ void MaxwellDMA::CopyBlockLinearToPitch() { UnswizzleSubrect(write_buffer, read_buffer, bytes_per_pixel, width, height, depth, x_offset, src_params.origin.y, x_elements, regs.line_count, block_height, block_depth, - regs.pitch_out); + abs_pitch_out); memory_manager.WriteBlockCached(regs.offset_out, write_buffer.data(), dst_size); } diff --git a/src/video_core/host1x/codecs/h264.cpp b/src/video_core/host1x/codecs/h264.cpp index 6ce179167..ce827eb6c 100644 --- a/src/video_core/host1x/codecs/h264.cpp +++ b/src/video_core/host1x/codecs/h264.cpp @@ -4,6 +4,7 @@ #include #include +#include "common/scratch_buffer.h" #include "common/settings.h" #include "video_core/host1x/codecs/h264.h" #include "video_core/host1x/host1x.h" @@ -188,7 +189,8 @@ void H264BitWriter::WriteBit(bool state) { } void H264BitWriter::WriteScalingList(std::span list, s32 start, s32 count) { - std::vector scan(count); + static Common::ScratchBuffer scan{}; + scan.resize_destructive(count); if (count == 16) { std::memcpy(scan.data(), zig_zag_scan.data(), scan.size()); } else { diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp index b2f7e160a..45141e488 100644 --- a/src/video_core/memory_manager.cpp +++ b/src/video_core/memory_manager.cpp @@ -587,7 +587,7 @@ void MemoryManager::InvalidateRegion(GPUVAddr gpu_addr, size_t size, void MemoryManager::CopyBlock(GPUVAddr gpu_dest_addr, GPUVAddr gpu_src_addr, std::size_t size, VideoCommon::CacheType which) { - std::vector tmp_buffer(size); + tmp_buffer.resize_destructive(size); ReadBlock(gpu_src_addr, tmp_buffer.data(), size, which); // The output block must be flushed in case it has data modified from the GPU. @@ -670,9 +670,9 @@ bool MemoryManager::IsFullyMappedRange(GPUVAddr gpu_addr, std::size_t size) cons return result; } -std::vector> MemoryManager::GetSubmappedRange( - GPUVAddr gpu_addr, std::size_t size) const { - std::vector> result{}; +boost::container::small_vector, 32> +MemoryManager::GetSubmappedRange(GPUVAddr gpu_addr, std::size_t size) const { + boost::container::small_vector, 32> result{}; GetSubmappedRangeImpl(gpu_addr, size, result); return result; } @@ -680,8 +680,9 @@ std::vector> MemoryManager::GetSubmappedRange( template void MemoryManager::GetSubmappedRangeImpl( GPUVAddr gpu_addr, std::size_t size, - std::vector, std::size_t>>& - result) const { + boost::container::small_vector< + std::pair, std::size_t>, 32>& result) + const { std::optional, std::size_t>> last_segment{}; std::optional old_page_addr{}; diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h index 794535122..4202c26ff 100644 --- a/src/video_core/memory_manager.h +++ b/src/video_core/memory_manager.h @@ -8,10 +8,12 @@ #include #include #include +#include #include "common/common_types.h" #include "common/multi_level_page_table.h" #include "common/range_map.h" +#include "common/scratch_buffer.h" #include "common/virtual_buffer.h" #include "video_core/cache_types.h" #include "video_core/pte_kind.h" @@ -107,8 +109,8 @@ public: * if the region is continuous, a single pair will be returned. If it's unmapped, an empty * vector will be returned; */ - std::vector> GetSubmappedRange(GPUVAddr gpu_addr, - std::size_t size) const; + boost::container::small_vector, 32> GetSubmappedRange( + GPUVAddr gpu_addr, std::size_t size) const; GPUVAddr Map(GPUVAddr gpu_addr, VAddr cpu_addr, std::size_t size, PTEKind kind = PTEKind::INVALID, bool is_big_pages = true); @@ -165,7 +167,8 @@ private: template void GetSubmappedRangeImpl( GPUVAddr gpu_addr, std::size_t size, - std::vector, std::size_t>>& + boost::container::small_vector< + std::pair, std::size_t>, 32>& result) const; Core::System& system; @@ -215,8 +218,8 @@ private: Common::VirtualBuffer big_page_table_cpu; std::vector big_page_continuous; - std::vector> page_stash{}; - std::vector> page_stash2{}; + boost::container::small_vector, 32> page_stash{}; + boost::container::small_vector, 32> page_stash2{}; mutable std::mutex guard; @@ -226,6 +229,8 @@ private: std::unique_ptr accumulator; static std::atomic unique_identifier_generator; + + Common::ScratchBuffer tmp_buffer; }; } // namespace Tegra diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 3f077311e..0329ed820 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -85,7 +85,9 @@ Shader::RuntimeInfo MakeRuntimeInfo(const GraphicsPipelineKey& key, case Shader::Stage::VertexB: case Shader::Stage::Geometry: if (!use_assembly_shaders && key.xfb_enabled != 0) { - info.xfb_varyings = VideoCommon::MakeTransformFeedbackVaryings(key.xfb_state); + auto [varyings, count] = VideoCommon::MakeTransformFeedbackVaryings(key.xfb_state); + info.xfb_varyings = varyings; + info.xfb_count = count; } break; case Shader::Stage::TessellationEval: diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp index e30fcb1ed..f47301ad5 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp @@ -361,7 +361,7 @@ void BufferCacheRuntime::CopyBuffer(VkBuffer dst_buffer, VkBuffer src_buffer, .dstAccessMask = VK_ACCESS_MEMORY_READ_BIT | VK_ACCESS_MEMORY_WRITE_BIT, }; // Measuring a popular game, this number never exceeds the specified size once data is warmed up - boost::container::small_vector vk_copies(copies.size()); + boost::container::small_vector vk_copies(copies.size()); std::ranges::transform(copies, vk_copies.begin(), MakeBufferCopy); scheduler.RequestOutsideRenderPassOperationContext(); scheduler.Record([src_buffer, dst_buffer, vk_copies, barrier](vk::CommandBuffer cmdbuf) { diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp index a2cfb2105..9f316113c 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp @@ -167,7 +167,10 @@ Shader::RuntimeInfo MakeRuntimeInfo(std::span program info.fixed_state_point_size = point_size; } if (key.state.xfb_enabled) { - info.xfb_varyings = VideoCommon::MakeTransformFeedbackVaryings(key.state.xfb_state); + auto [varyings, count] = + VideoCommon::MakeTransformFeedbackVaryings(key.state.xfb_state); + info.xfb_varyings = varyings; + info.xfb_count = count; } info.convert_depth_mode = gl_ndc; } @@ -214,7 +217,10 @@ Shader::RuntimeInfo MakeRuntimeInfo(std::span program info.fixed_state_point_size = point_size; } if (key.state.xfb_enabled != 0) { - info.xfb_varyings = VideoCommon::MakeTransformFeedbackVaryings(key.state.xfb_state); + auto [varyings, count] = + VideoCommon::MakeTransformFeedbackVaryings(key.state.xfb_state); + info.xfb_varyings = varyings; + info.xfb_count = count; } info.convert_depth_mode = gl_ndc; break; diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index f025f618b..f3cef09dd 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -330,9 +330,9 @@ constexpr VkBorderColor ConvertBorderColor(const std::array& color) { }; } -[[maybe_unused]] [[nodiscard]] std::vector TransformBufferCopies( - std::span copies, size_t buffer_offset) { - std::vector result(copies.size()); +[[maybe_unused]] [[nodiscard]] boost::container::small_vector +TransformBufferCopies(std::span copies, size_t buffer_offset) { + boost::container::small_vector result(copies.size()); std::ranges::transform( copies, result.begin(), [buffer_offset](const VideoCommon::BufferCopy& copy) { return VkBufferCopy{ @@ -344,7 +344,7 @@ constexpr VkBorderColor ConvertBorderColor(const std::array& color) { return result; } -[[nodiscard]] std::vector TransformBufferImageCopies( +[[nodiscard]] boost::container::small_vector TransformBufferImageCopies( std::span copies, size_t buffer_offset, VkImageAspectFlags aspect_mask) { struct Maker { VkBufferImageCopy operator()(const BufferImageCopy& copy) const { @@ -377,14 +377,14 @@ constexpr VkBorderColor ConvertBorderColor(const std::array& color) { VkImageAspectFlags aspect_mask; }; if (aspect_mask == (VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT)) { - std::vector result(copies.size() * 2); + boost::container::small_vector result(copies.size() * 2); std::ranges::transform(copies, result.begin(), Maker{buffer_offset, VK_IMAGE_ASPECT_DEPTH_BIT}); std::ranges::transform(copies, result.begin() + copies.size(), Maker{buffer_offset, VK_IMAGE_ASPECT_STENCIL_BIT}); return result; } else { - std::vector result(copies.size()); + boost::container::small_vector result(copies.size()); std::ranges::transform(copies, result.begin(), Maker{buffer_offset, aspect_mask}); return result; } @@ -867,8 +867,8 @@ void TextureCacheRuntime::BarrierFeedbackLoop() { void TextureCacheRuntime::ReinterpretImage(Image& dst, Image& src, std::span copies) { - std::vector vk_in_copies(copies.size()); - std::vector vk_out_copies(copies.size()); + boost::container::small_vector vk_in_copies(copies.size()); + boost::container::small_vector vk_out_copies(copies.size()); const VkImageAspectFlags src_aspect_mask = src.AspectMask(); const VkImageAspectFlags dst_aspect_mask = dst.AspectMask(); @@ -1157,7 +1157,7 @@ void TextureCacheRuntime::ConvertImage(Framebuffer* dst, ImageView& dst_view, Im void TextureCacheRuntime::CopyImage(Image& dst, Image& src, std::span copies) { - std::vector vk_copies(copies.size()); + boost::container::small_vector vk_copies(copies.size()); const VkImageAspectFlags aspect_mask = dst.AspectMask(); ASSERT(aspect_mask == src.AspectMask()); @@ -1332,7 +1332,7 @@ void Image::UploadMemory(VkBuffer buffer, VkDeviceSize offset, ScaleDown(true); } scheduler->RequestOutsideRenderPassOperationContext(); - std::vector vk_copies = TransformBufferImageCopies(copies, offset, aspect_mask); + auto vk_copies = TransformBufferImageCopies(copies, offset, aspect_mask); const VkBuffer src_buffer = buffer; const VkImage vk_image = *original_image; const VkImageAspectFlags vk_aspect_mask = aspect_mask; @@ -1367,8 +1367,9 @@ void Image::DownloadMemory(std::span buffers_span, std::span buffers_vector{}; - boost::container::small_vector, 1> vk_copies; + boost::container::small_vector buffers_vector{}; + boost::container::small_vector, 8> + vk_copies; for (size_t index = 0; index < buffers_span.size(); index++) { buffers_vector.emplace_back(buffers_span[index]); vk_copies.emplace_back( @@ -1858,7 +1859,7 @@ Framebuffer::~Framebuffer() = default; void Framebuffer::CreateFramebuffer(TextureCacheRuntime& runtime, std::span color_buffers, ImageView* depth_buffer, bool is_rescaled) { - std::vector attachments; + boost::container::small_vector attachments; RenderPassKey renderpass_key{}; s32 num_layers = 1; diff --git a/src/video_core/shader_cache.cpp b/src/video_core/shader_cache.cpp index c5213875b..4db948b6d 100644 --- a/src/video_core/shader_cache.cpp +++ b/src/video_core/shader_cache.cpp @@ -151,11 +151,9 @@ void ShaderCache::RemovePendingShaders() { marked_for_removal.erase(std::unique(marked_for_removal.begin(), marked_for_removal.end()), marked_for_removal.end()); - std::vector removed_shaders; - removed_shaders.reserve(marked_for_removal.size()); + boost::container::small_vector removed_shaders; std::scoped_lock lock{lookup_mutex}; - for (Entry* const entry : marked_for_removal) { removed_shaders.push_back(entry->data); diff --git a/src/video_core/texture_cache/image_base.h b/src/video_core/texture_cache/image_base.h index 1b8a17ee8..55d49d017 100644 --- a/src/video_core/texture_cache/image_base.h +++ b/src/video_core/texture_cache/image_base.h @@ -6,6 +6,7 @@ #include #include #include +#include #include "common/common_funcs.h" #include "common/common_types.h" @@ -108,8 +109,8 @@ struct ImageBase { std::vector image_view_infos; std::vector image_view_ids; - std::vector slice_offsets; - std::vector slice_subresources; + boost::container::small_vector slice_offsets; + boost::container::small_vector slice_subresources; std::vector aliased_images; std::vector overlapping_images; diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index d58bb69ff..d3f03a995 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -526,7 +526,7 @@ void TextureCache

::WriteMemory(VAddr cpu_addr, size_t size) { template void TextureCache

::DownloadMemory(VAddr cpu_addr, size_t size) { - std::vector images; + boost::container::small_vector images; ForEachImageInRegion(cpu_addr, size, [&images](ImageId image_id, ImageBase& image) { if (!image.IsSafeDownload()) { return; @@ -579,7 +579,7 @@ std::optional TextureCache

::GetFlushArea(V template void TextureCache

::UnmapMemory(VAddr cpu_addr, size_t size) { - std::vector deleted_images; + boost::container::small_vector deleted_images; ForEachImageInRegion(cpu_addr, size, [&](ImageId id, Image&) { deleted_images.push_back(id); }); for (const ImageId id : deleted_images) { Image& image = slot_images[id]; @@ -593,7 +593,7 @@ void TextureCache

::UnmapMemory(VAddr cpu_addr, size_t size) { template void TextureCache

::UnmapGPUMemory(size_t as_id, GPUVAddr gpu_addr, size_t size) { - std::vector deleted_images; + boost::container::small_vector deleted_images; ForEachImageInRegionGPU(as_id, gpu_addr, size, [&](ImageId id, Image&) { deleted_images.push_back(id); }); for (const ImageId id : deleted_images) { @@ -1101,7 +1101,7 @@ ImageId TextureCache

::FindImage(const ImageInfo& info, GPUVAddr gpu_addr, const bool native_bgr = runtime.HasNativeBgr(); const bool flexible_formats = True(options & RelaxedOptions::Format); ImageId image_id{}; - boost::container::small_vector image_ids; + boost::container::small_vector image_ids; const auto lambda = [&](ImageId existing_image_id, ImageBase& existing_image) { if (True(existing_image.flags & ImageFlagBits::Remapped)) { return false; @@ -1622,7 +1622,7 @@ ImageId TextureCache

::FindDMAImage(const ImageInfo& info, GPUVAddr gpu_addr) } } ImageId image_id{}; - boost::container::small_vector image_ids; + boost::container::small_vector image_ids; const auto lambda = [&](ImageId existing_image_id, ImageBase& existing_image) { if (True(existing_image.flags & ImageFlagBits::Remapped)) { return false; @@ -1942,7 +1942,7 @@ void TextureCache

::RegisterImage(ImageId image_id) { image.map_view_id = map_id; return; } - std::vector sparse_maps{}; + boost::container::small_vector sparse_maps; ForEachSparseSegment( image, [this, image_id, &sparse_maps](GPUVAddr gpu_addr, VAddr cpu_addr, size_t size) { auto map_id = slot_map_views.insert(gpu_addr, cpu_addr, size, image_id); @@ -2217,7 +2217,7 @@ void TextureCache

::MarkModification(ImageBase& image) noexcept { template void TextureCache

::SynchronizeAliases(ImageId image_id) { - boost::container::small_vector aliased_images; + boost::container::small_vector aliased_images; Image& image = slot_images[image_id]; bool any_rescaled = True(image.flags & ImageFlagBits::Rescaled); bool any_modified = True(image.flags & ImageFlagBits::GpuModified); diff --git a/src/video_core/texture_cache/texture_cache_base.h b/src/video_core/texture_cache/texture_cache_base.h index 44232b961..e9ec91265 100644 --- a/src/video_core/texture_cache/texture_cache_base.h +++ b/src/video_core/texture_cache/texture_cache_base.h @@ -56,7 +56,7 @@ struct ImageViewInOut { struct AsyncDecodeContext { ImageId image_id; Common::ScratchBuffer decoded_data; - std::vector copies; + boost::container::small_vector copies; std::mutex mutex; std::atomic_bool complete; }; @@ -429,7 +429,7 @@ private: std::unordered_map, Common::IdentityHash> page_table; std::unordered_map, Common::IdentityHash> sparse_page_table; - std::unordered_map> sparse_views; + std::unordered_map> sparse_views; VAddr virtual_invalid_space{}; diff --git a/src/video_core/texture_cache/util.cpp b/src/video_core/texture_cache/util.cpp index 95a5b47d8..f781cb7a0 100644 --- a/src/video_core/texture_cache/util.cpp +++ b/src/video_core/texture_cache/util.cpp @@ -329,13 +329,13 @@ template [[nodiscard]] std::optional ResolveOverlapRightAddress3D( const ImageInfo& new_info, GPUVAddr gpu_addr, const ImageBase& overlap, bool strict_size) { - const std::vector slice_offsets = CalculateSliceOffsets(new_info); + const auto slice_offsets = CalculateSliceOffsets(new_info); const u32 diff = static_cast(overlap.gpu_addr - gpu_addr); const auto it = std::ranges::find(slice_offsets, diff); if (it == slice_offsets.end()) { return std::nullopt; } - const std::vector subresources = CalculateSliceSubresources(new_info); + const auto subresources = CalculateSliceSubresources(new_info); const SubresourceBase base = subresources[std::distance(slice_offsets.begin(), it)]; const ImageInfo& info = overlap.info; if (!IsBlockLinearSizeCompatible(new_info, info, base.level, 0, strict_size)) { @@ -655,9 +655,9 @@ LevelArray CalculateMipLevelSizes(const ImageInfo& info) noexcept { return sizes; } -std::vector CalculateSliceOffsets(const ImageInfo& info) { +boost::container::small_vector CalculateSliceOffsets(const ImageInfo& info) { ASSERT(info.type == ImageType::e3D); - std::vector offsets; + boost::container::small_vector offsets; offsets.reserve(NumSlices(info)); const LevelInfo level_info = MakeLevelInfo(info); @@ -679,9 +679,10 @@ std::vector CalculateSliceOffsets(const ImageInfo& info) { return offsets; } -std::vector CalculateSliceSubresources(const ImageInfo& info) { +boost::container::small_vector CalculateSliceSubresources( + const ImageInfo& info) { ASSERT(info.type == ImageType::e3D); - std::vector subresources; + boost::container::small_vector subresources; subresources.reserve(NumSlices(info)); for (s32 level = 0; level < info.resources.levels; ++level) { const s32 depth = AdjustMipSize(info.size.depth, level); @@ -723,8 +724,10 @@ ImageViewType RenderTargetImageViewType(const ImageInfo& info) noexcept { } } -std::vector MakeShrinkImageCopies(const ImageInfo& dst, const ImageInfo& src, - SubresourceBase base, u32 up_scale, u32 down_shift) { +boost::container::small_vector MakeShrinkImageCopies(const ImageInfo& dst, + const ImageInfo& src, + SubresourceBase base, + u32 up_scale, u32 down_shift) { ASSERT(dst.resources.levels >= src.resources.levels); const bool is_dst_3d = dst.type == ImageType::e3D; @@ -733,7 +736,7 @@ std::vector MakeShrinkImageCopies(const ImageInfo& dst, const ImageIn ASSERT(src.resources.levels == 1); } const bool both_2d{src.type == ImageType::e2D && dst.type == ImageType::e2D}; - std::vector copies; + boost::container::small_vector copies; copies.reserve(src.resources.levels); for (s32 level = 0; level < src.resources.levels; ++level) { ImageCopy& copy = copies.emplace_back(); @@ -770,9 +773,10 @@ std::vector MakeShrinkImageCopies(const ImageInfo& dst, const ImageIn return copies; } -std::vector MakeReinterpretImageCopies(const ImageInfo& src, u32 up_scale, - u32 down_shift) { - std::vector copies; +boost::container::small_vector MakeReinterpretImageCopies(const ImageInfo& src, + u32 up_scale, + u32 down_shift) { + boost::container::small_vector copies; copies.reserve(src.resources.levels); const bool is_3d = src.type == ImageType::e3D; for (s32 level = 0; level < src.resources.levels; ++level) { @@ -824,9 +828,11 @@ bool IsValidEntry(const Tegra::MemoryManager& gpu_memory, const TICEntry& config return gpu_memory.GpuToCpuAddress(address, guest_size_bytes).has_value(); } -std::vector UnswizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, - const ImageInfo& info, std::span input, - std::span output) { +boost::container::small_vector UnswizzleImage(Tegra::MemoryManager& gpu_memory, + GPUVAddr gpu_addr, + const ImageInfo& info, + std::span input, + std::span output) { const size_t guest_size_bytes = input.size_bytes(); const u32 bpp_log2 = BytesPerBlockLog2(info.format); const Extent3D size = info.size; @@ -861,7 +867,7 @@ std::vector UnswizzleImage(Tegra::MemoryManager& gpu_memory, GP info.tile_width_spacing); size_t guest_offset = 0; u32 host_offset = 0; - std::vector copies(num_levels); + boost::container::small_vector copies(num_levels); for (s32 level = 0; level < num_levels; ++level) { const Extent3D level_size = AdjustMipSize(size, level); @@ -978,7 +984,7 @@ void ConvertImage(std::span input, const ImageInfo& info, std::span FullDownloadCopies(const ImageInfo& info) { +boost::container::small_vector FullDownloadCopies(const ImageInfo& info) { const Extent3D size = info.size; const u32 bytes_per_block = BytesPerBlock(info.format); if (info.type == ImageType::Linear) { @@ -1006,7 +1012,7 @@ std::vector FullDownloadCopies(const ImageInfo& info) { u32 host_offset = 0; - std::vector copies(num_levels); + boost::container::small_vector copies(num_levels); for (s32 level = 0; level < num_levels; ++level) { const Extent3D level_size = AdjustMipSize(size, level); const u32 num_blocks_per_layer = NumBlocks(level_size, tile_size); @@ -1042,10 +1048,10 @@ Extent3D MipBlockSize(const ImageInfo& info, u32 level) { return AdjustMipBlockSize(num_tiles, level_info.block, level); } -std::vector FullUploadSwizzles(const ImageInfo& info) { +boost::container::small_vector FullUploadSwizzles(const ImageInfo& info) { const Extent2D tile_size = DefaultBlockSize(info.format); if (info.type == ImageType::Linear) { - return std::vector{SwizzleParameters{ + return {SwizzleParameters{ .num_tiles = AdjustTileSize(info.size, tile_size), .block = {}, .buffer_offset = 0, @@ -1057,7 +1063,7 @@ std::vector FullUploadSwizzles(const ImageInfo& info) { const s32 num_levels = info.resources.levels; u32 guest_offset = 0; - std::vector params(num_levels); + boost::container::small_vector params(num_levels); for (s32 level = 0; level < num_levels; ++level) { const Extent3D level_size = AdjustMipSize(size, level); const Extent3D num_tiles = AdjustTileSize(level_size, tile_size); diff --git a/src/video_core/texture_cache/util.h b/src/video_core/texture_cache/util.h index 84aa6880d..ab45a43c4 100644 --- a/src/video_core/texture_cache/util.h +++ b/src/video_core/texture_cache/util.h @@ -5,6 +5,7 @@ #include #include +#include #include "common/common_types.h" #include "common/scratch_buffer.h" @@ -40,9 +41,10 @@ struct OverlapResult { [[nodiscard]] LevelArray CalculateMipLevelSizes(const ImageInfo& info) noexcept; -[[nodiscard]] std::vector CalculateSliceOffsets(const ImageInfo& info); +[[nodiscard]] boost::container::small_vector CalculateSliceOffsets(const ImageInfo& info); -[[nodiscard]] std::vector CalculateSliceSubresources(const ImageInfo& info); +[[nodiscard]] boost::container::small_vector CalculateSliceSubresources( + const ImageInfo& info); [[nodiscard]] u32 CalculateLevelStrideAlignment(const ImageInfo& info, u32 level); @@ -51,21 +53,18 @@ struct OverlapResult { [[nodiscard]] ImageViewType RenderTargetImageViewType(const ImageInfo& info) noexcept; -[[nodiscard]] std::vector MakeShrinkImageCopies(const ImageInfo& dst, - const ImageInfo& src, - SubresourceBase base, u32 up_scale = 1, - u32 down_shift = 0); +[[nodiscard]] boost::container::small_vector MakeShrinkImageCopies( + const ImageInfo& dst, const ImageInfo& src, SubresourceBase base, u32 up_scale = 1, + u32 down_shift = 0); -[[nodiscard]] std::vector MakeReinterpretImageCopies(const ImageInfo& src, - u32 up_scale = 1, - u32 down_shift = 0); +[[nodiscard]] boost::container::small_vector MakeReinterpretImageCopies( + const ImageInfo& src, u32 up_scale = 1, u32 down_shift = 0); [[nodiscard]] bool IsValidEntry(const Tegra::MemoryManager& gpu_memory, const TICEntry& config); -[[nodiscard]] std::vector UnswizzleImage(Tegra::MemoryManager& gpu_memory, - GPUVAddr gpu_addr, const ImageInfo& info, - std::span input, - std::span output); +[[nodiscard]] boost::container::small_vector UnswizzleImage( + Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const ImageInfo& info, + std::span input, std::span output); [[nodiscard]] BufferCopy UploadBufferCopy(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const ImageBase& image, std::span output); @@ -73,13 +72,15 @@ struct OverlapResult { void ConvertImage(std::span input, const ImageInfo& info, std::span output, std::span copies); -[[nodiscard]] std::vector FullDownloadCopies(const ImageInfo& info); +[[nodiscard]] boost::container::small_vector FullDownloadCopies( + const ImageInfo& info); [[nodiscard]] Extent3D MipSize(Extent3D size, u32 level); [[nodiscard]] Extent3D MipBlockSize(const ImageInfo& info, u32 level); -[[nodiscard]] std::vector FullUploadSwizzles(const ImageInfo& info); +[[nodiscard]] boost::container::small_vector FullUploadSwizzles( + const ImageInfo& info); void SwizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const ImageInfo& info, std::span copies, std::span memory, diff --git a/src/video_core/transform_feedback.cpp b/src/video_core/transform_feedback.cpp index 155599316..1f353d2df 100644 --- a/src/video_core/transform_feedback.cpp +++ b/src/video_core/transform_feedback.cpp @@ -13,7 +13,7 @@ namespace VideoCommon { -std::vector MakeTransformFeedbackVaryings( +std::pair, u32> MakeTransformFeedbackVaryings( const TransformFeedbackState& state) { static constexpr std::array VECTORS{ 28U, // gl_Position @@ -62,7 +62,8 @@ std::vector MakeTransformFeedbackVaryings( 216U, // gl_TexCoord[6] 220U, // gl_TexCoord[7] }; - std::vector xfb(256); + std::array xfb{}; + u32 count{0}; for (size_t buffer = 0; buffer < state.layouts.size(); ++buffer) { const auto& locations = state.varyings[buffer]; const auto& layout = state.layouts[buffer]; @@ -103,11 +104,12 @@ std::vector MakeTransformFeedbackVaryings( } } xfb[attribute] = varying; + count = std::max(count, attribute); highest = std::max(highest, (base_offset + varying.components) * 4); } UNIMPLEMENTED_IF(highest != layout.stride); } - return xfb; + return {xfb, count + 1}; } } // namespace VideoCommon diff --git a/src/video_core/transform_feedback.h b/src/video_core/transform_feedback.h index d13eb16c3..401b1352a 100644 --- a/src/video_core/transform_feedback.h +++ b/src/video_core/transform_feedback.h @@ -24,7 +24,7 @@ struct TransformFeedbackState { varyings; }; -std::vector MakeTransformFeedbackVaryings( +std::pair, u32> MakeTransformFeedbackVaryings( const TransformFeedbackState& state); } // namespace VideoCommon diff --git a/src/video_core/vulkan_common/vulkan_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp index fa9cde75b..b11abe311 100644 --- a/src/video_core/vulkan_common/vulkan_device.cpp +++ b/src/video_core/vulkan_common/vulkan_device.cpp @@ -316,6 +316,7 @@ NvidiaArchitecture GetNvidiaArchitecture(vk::PhysicalDevice physical, std::vector ExtensionListForVulkan( const std::set>& extensions) { std::vector output; + output.reserve(extensions.size()); for (const auto& extension : extensions) { output.push_back(extension.c_str()); }