diff options
Diffstat (limited to 'src')
49 files changed, 4209 insertions, 2113 deletions
diff --git a/src/common/input.h b/src/common/input.h index bfa0639f5..b533f3844 100644 --- a/src/common/input.h +++ b/src/common/input.h @@ -277,8 +277,9 @@ struct CallbackStatus { BodyColorStatus color_status{}; BatteryStatus battery_status{}; VibrationStatus vibration_status{}; - CameraStatus camera_status{}; - NfcStatus nfc_status{}; + CameraFormat camera_status{CameraFormat::None}; + NfcState nfc_status{NfcState::Unknown}; + std::vector<u8> raw_data{}; }; // Triggered once every input change diff --git a/src/common/settings.h b/src/common/settings.h index d2452c93b..0eb98939c 100644 --- a/src/common/settings.h +++ b/src/common/settings.h @@ -431,7 +431,7 @@ struct Values { FullscreenMode::Exclusive, #endif FullscreenMode::Borderless, FullscreenMode::Exclusive, "fullscreen_mode"}; - SwitchableSetting<int, true> aspect_ratio{0, 0, 3, "aspect_ratio"}; + SwitchableSetting<int, true> aspect_ratio{0, 0, 4, "aspect_ratio"}; SwitchableSetting<int, true> max_anisotropy{0, 0, 5, "max_anisotropy"}; SwitchableSetting<bool> use_speed_limit{true, "use_speed_limit"}; SwitchableSetting<u16, true> speed_limit{100, 0, 9999, "speed_limit"}; diff --git a/src/core/core_timing.cpp b/src/core/core_timing.cpp index 6c0fcb7b5..2678ce532 100644 --- a/src/core/core_timing.cpp +++ b/src/core/core_timing.cpp @@ -270,6 +270,7 @@ void CoreTiming::ThreadLoop() { // There are more events left in the queue, wait until the next event. const auto wait_time = *next_time - GetGlobalTimeNs().count(); if (wait_time > 0) { +#ifdef _WIN32 // Assume a timer resolution of 1ms. static constexpr s64 TimerResolutionNS = 1000000; @@ -287,6 +288,9 @@ void CoreTiming::ThreadLoop() { if (event.IsSet()) { event.Reset(); } +#else + event.WaitFor(std::chrono::nanoseconds(wait_time)); +#endif } } else { // Queue is empty, wait until another event is scheduled and signals us to continue. diff --git a/src/core/frontend/framebuffer_layout.cpp b/src/core/frontend/framebuffer_layout.cpp index 90dd68ff1..b4081fc39 100644 --- a/src/core/frontend/framebuffer_layout.cpp +++ b/src/core/frontend/framebuffer_layout.cpp @@ -67,6 +67,8 @@ float EmulationAspectRatio(AspectRatio aspect, float window_aspect_ratio) { return 3.0f / 4.0f; case AspectRatio::R21_9: return 9.0f / 21.0f; + case AspectRatio::R16_10: + return 10.0f / 16.0f; case AspectRatio::StretchToWindow: return window_aspect_ratio; default: diff --git a/src/core/frontend/framebuffer_layout.h b/src/core/frontend/framebuffer_layout.h index 1561d994e..94683b30f 100644 --- a/src/core/frontend/framebuffer_layout.h +++ b/src/core/frontend/framebuffer_layout.h @@ -27,6 +27,7 @@ enum class AspectRatio { Default, R4_3, R21_9, + R16_10, StretchToWindow, }; diff --git a/src/core/hid/input_converter.cpp b/src/core/hid/input_converter.cpp index fe9915abe..5d8b75b50 100644 --- a/src/core/hid/input_converter.cpp +++ b/src/core/hid/input_converter.cpp @@ -277,7 +277,10 @@ Common::Input::CameraStatus TransformToCamera(const Common::Input::CallbackStatu Common::Input::CameraStatus camera{}; switch (callback.type) { case Common::Input::InputType::IrSensor: - camera = callback.camera_status; + camera = { + .format = callback.camera_status, + .data = callback.raw_data, + }; break; default: LOG_ERROR(Input, "Conversion from type {} to camera not implemented", callback.type); @@ -291,7 +294,10 @@ Common::Input::NfcStatus TransformToNfc(const Common::Input::CallbackStatus& cal Common::Input::NfcStatus nfc{}; switch (callback.type) { case Common::Input::InputType::Nfc: - return callback.nfc_status; + nfc = { + .state = callback.nfc_status, + .data = callback.raw_data, + }; break; default: LOG_ERROR(Input, "Conversion from type {} to NFC not implemented", callback.type); diff --git a/src/core/hle/result.h b/src/core/hle/result.h index 47a1b829b..d67e68bae 100644 --- a/src/core/hle/result.h +++ b/src/core/hle/result.h @@ -5,6 +5,7 @@ #include "common/assert.h" #include "common/bit_field.h" +#include "common/common_funcs.h" #include "common/common_types.h" #include "common/expected.h" @@ -130,6 +131,10 @@ union Result { [[nodiscard]] constexpr bool IsError() const { return !IsSuccess(); } + + [[nodiscard]] constexpr bool IsFailure() const { + return !IsSuccess(); + } }; static_assert(std::is_trivial_v<Result>); @@ -349,10 +354,109 @@ private: } \ } while (false) -#define R_SUCCEEDED(res) (res.IsSuccess()) +#define R_SUCCEEDED(res) (static_cast<Result>(res).IsSuccess()) +#define R_FAILED(res) (static_cast<Result>(res).IsFailure()) -/// Evaluates a boolean expression, and succeeds if that expression is true. -#define R_SUCCEED_IF(expr) R_UNLESS(!(expr), ResultSuccess) +namespace ResultImpl { +template <auto EvaluateResult, class F> +class ScopedResultGuard { + YUZU_NON_COPYABLE(ScopedResultGuard); + YUZU_NON_MOVEABLE(ScopedResultGuard); + +private: + Result& m_ref; + F m_f; + +public: + constexpr ScopedResultGuard(Result& ref, F f) : m_ref(ref), m_f(std::move(f)) {} + constexpr ~ScopedResultGuard() { + if (EvaluateResult(m_ref)) { + m_f(); + } + } +}; + +template <auto EvaluateResult> +class ResultReferenceForScopedResultGuard { +private: + Result& m_ref; + +public: + constexpr ResultReferenceForScopedResultGuard(Result& r) : m_ref(r) {} + constexpr operator Result&() const { + return m_ref; + } +}; + +template <auto EvaluateResult, typename F> +constexpr ScopedResultGuard<EvaluateResult, F> operator+( + ResultReferenceForScopedResultGuard<EvaluateResult> ref, F&& f) { + return ScopedResultGuard<EvaluateResult, F>(static_cast<Result&>(ref), std::forward<F>(f)); +} + +constexpr bool EvaluateResultSuccess(const Result& r) { + return R_SUCCEEDED(r); +} +constexpr bool EvaluateResultFailure(const Result& r) { + return R_FAILED(r); +} + +template <typename T> +constexpr void UpdateCurrentResultReference(T result_reference, Result result) = delete; +// Intentionally not defined + +template <> +constexpr void UpdateCurrentResultReference<Result&>(Result& result_reference, Result result) { + result_reference = result; +} + +template <> +constexpr void UpdateCurrentResultReference<const Result>(Result result_reference, Result result) {} +} // namespace ResultImpl + +#define DECLARE_CURRENT_RESULT_REFERENCE_AND_STORAGE(COUNTER_VALUE) \ + [[maybe_unused]] constexpr bool HasPrevRef_##COUNTER_VALUE = \ + std::same_as<decltype(__TmpCurrentResultReference), Result&>; \ + [[maybe_unused]] auto& PrevRef_##COUNTER_VALUE = __TmpCurrentResultReference; \ + [[maybe_unused]] Result __tmp_result_##COUNTER_VALUE = ResultSuccess; \ + Result& __TmpCurrentResultReference = \ + HasPrevRef_##COUNTER_VALUE ? PrevRef_##COUNTER_VALUE : __tmp_result_##COUNTER_VALUE + +#define ON_RESULT_RETURN_IMPL(...) \ + static_assert(std::same_as<decltype(__TmpCurrentResultReference), Result&>); \ + auto RESULT_GUARD_STATE_##__COUNTER__ = \ + ResultImpl::ResultReferenceForScopedResultGuard<__VA_ARGS__>( \ + __TmpCurrentResultReference) + \ + [&]() + +#define ON_RESULT_FAILURE_2 ON_RESULT_RETURN_IMPL(ResultImpl::EvaluateResultFailure) + +#define ON_RESULT_FAILURE \ + DECLARE_CURRENT_RESULT_REFERENCE_AND_STORAGE(__COUNTER__); \ + ON_RESULT_FAILURE_2 + +#define ON_RESULT_SUCCESS_2 ON_RESULT_RETURN_IMPL(ResultImpl::EvaluateResultSuccess) + +#define ON_RESULT_SUCCESS \ + DECLARE_CURRENT_RESULT_REFERENCE_AND_STORAGE(__COUNTER__); \ + ON_RESULT_SUCCESS_2 + +constexpr inline Result __TmpCurrentResultReference = ResultSuccess; + +/// Returns a result. +#define R_RETURN(res_expr) \ + { \ + const Result _tmp_r_throw_rc = (res_expr); \ + ResultImpl::UpdateCurrentResultReference<decltype(__TmpCurrentResultReference)>( \ + __TmpCurrentResultReference, _tmp_r_throw_rc); \ + return _tmp_r_throw_rc; \ + } + +/// Returns ResultSuccess() +#define R_SUCCEED() R_RETURN(ResultSuccess) + +/// Throws a result. +#define R_THROW(res_expr) R_RETURN(res_expr) /// Evaluates a boolean expression, and returns a result unless that expression is true. #define R_UNLESS(expr, res) \ @@ -361,7 +465,7 @@ private: if (res.IsError()) { \ LOG_ERROR(Kernel, "Failed with result: {}", res.raw); \ } \ - return res; \ + R_THROW(res); \ } \ } @@ -369,7 +473,10 @@ private: #define R_TRY(res_expr) \ { \ const auto _tmp_r_try_rc = (res_expr); \ - if (_tmp_r_try_rc.IsError()) { \ - return _tmp_r_try_rc; \ + if (R_FAILED(_tmp_r_try_rc)) { \ + R_THROW(_tmp_r_try_rc); \ } \ } + +/// Evaluates a boolean expression, and succeeds if that expression is true. +#define R_SUCCEED_IF(expr) R_UNLESS(!(expr), ResultSuccess) diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp index e23eae36a..c08274ef9 100644 --- a/src/core/hle/service/filesystem/fsp_srv.cpp +++ b/src/core/hle/service/filesystem/fsp_srv.cpp @@ -707,7 +707,7 @@ FSP_SRV::FSP_SRV(Core::System& system_) {31, nullptr, "OpenGameCardFileSystem"}, {32, nullptr, "ExtendSaveDataFileSystem"}, {33, nullptr, "DeleteCacheStorage"}, - {34, nullptr, "GetCacheStorageSize"}, + {34, &FSP_SRV::GetCacheStorageSize, "GetCacheStorageSize"}, {35, nullptr, "CreateSaveDataFileSystemByHashSalt"}, {36, nullptr, "OpenHostFileSystemWithOption"}, {51, &FSP_SRV::OpenSaveDataFileSystem, "OpenSaveDataFileSystem"}, @@ -1107,6 +1107,18 @@ void FSP_SRV::GetProgramIndexForAccessLog(Kernel::HLERequestContext& ctx) { rb.Push(access_log_program_index); } +void FSP_SRV::GetCacheStorageSize(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto index{rp.Pop<s32>()}; + + LOG_WARNING(Service_FS, "(STUBBED) called with index={}", index); + + IPC::ResponseBuilder rb{ctx, 6}; + rb.Push(ResultSuccess); + rb.Push(s64{0}); + rb.Push(s64{0}); +} + class IMultiCommitManager final : public ServiceFramework<IMultiCommitManager> { public: explicit IMultiCommitManager(Core::System& system_) diff --git a/src/core/hle/service/filesystem/fsp_srv.h b/src/core/hle/service/filesystem/fsp_srv.h index 36f552e05..3d88b97f9 100644 --- a/src/core/hle/service/filesystem/fsp_srv.h +++ b/src/core/hle/service/filesystem/fsp_srv.h @@ -54,6 +54,7 @@ private: void OutputAccessLogToSdCard(Kernel::HLERequestContext& ctx); void GetProgramIndexForAccessLog(Kernel::HLERequestContext& ctx); void OpenMultiCommitManager(Kernel::HLERequestContext& ctx); + void GetCacheStorageSize(Kernel::HLERequestContext& ctx); FileSystemController& fsc; const FileSys::ContentProvider& content_provider; diff --git a/src/core/hle/service/friend/friend.cpp b/src/core/hle/service/friend/friend.cpp index e0db787fc..fad532115 100644 --- a/src/core/hle/service/friend/friend.cpp +++ b/src/core/hle/service/friend/friend.cpp @@ -26,7 +26,7 @@ public: {10101, &IFriendService::GetFriendList, "GetFriendList"}, {10102, nullptr, "UpdateFriendInfo"}, {10110, nullptr, "GetFriendProfileImage"}, - {10120, nullptr, "IsFriendListCacheAvailable"}, + {10120, &IFriendService::CheckFriendListAvailability, "CheckFriendListAvailability"}, {10121, nullptr, "EnsureFriendListAvailable"}, {10200, nullptr, "SendFriendRequestForApplication"}, {10211, nullptr, "AddFacedFriendRequestForApplication"}, @@ -194,6 +194,17 @@ private: // TODO(ogniK): Return a buffer of u64s which are the "NetworkServiceAccountId" } + void CheckFriendListAvailability(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto uuid{rp.PopRaw<Common::UUID>()}; + + LOG_WARNING(Service_Friend, "(STUBBED) called, uuid=0x{}", uuid.RawString()); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(ResultSuccess); + rb.Push(true); + } + KernelHelpers::ServiceContext service_context; Kernel::KEvent* completion_event; diff --git a/src/core/hle/service/nfp/nfp_types.h b/src/core/hle/service/nfp/nfp_types.h index 867ea2f36..c09f9ddb6 100644 --- a/src/core/hle/service/nfp/nfp_types.h +++ b/src/core/hle/service/nfp/nfp_types.h @@ -167,7 +167,7 @@ struct AmiiboDate { bool IsValidDate() const { const bool is_day_valid = GetDay() > 0 && GetDay() < 32; - const bool is_month_valid = GetMonth() >= 0 && GetMonth() < 13; + const bool is_month_valid = GetMonth() > 0 && GetMonth() < 13; const bool is_year_valid = GetYear() >= 2000; return is_year_valid && is_month_valid && is_day_valid; } diff --git a/src/input_common/input_poller.cpp b/src/input_common/input_poller.cpp index 75705b67e..ca33fb4eb 100644 --- a/src/input_common/input_poller.cpp +++ b/src/input_common/input_poller.cpp @@ -691,9 +691,12 @@ public: } void OnChange() { + const auto camera_status = GetStatus(); + const Common::Input::CallbackStatus status{ .type = Common::Input::InputType::IrSensor, - .camera_status = GetStatus(), + .camera_status = camera_status.format, + .raw_data = camera_status.data, }; TriggerOnChange(status); @@ -732,9 +735,12 @@ public: } void OnChange() { + const auto nfc_status = GetStatus(); + const Common::Input::CallbackStatus status{ .type = Common::Input::InputType::Nfc, - .nfc_status = GetStatus(), + .nfc_status = nfc_status.state, + .raw_data = nfc_status.data, }; TriggerOnChange(status); diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 8e26b3f95..2ba33543c 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -994,13 +994,13 @@ void BufferCache<P>::BindHostIndexBuffer() { const u32 size = index_buffer.size; SynchronizeBuffer(buffer, index_buffer.cpu_addr, size); if constexpr (HAS_FULL_INDEX_AND_PRIMITIVE_SUPPORT) { - const u32 new_offset = offset + maxwell3d->regs.index_array.first * - maxwell3d->regs.index_array.FormatSizeInBytes(); + const u32 new_offset = offset + maxwell3d->regs.index_buffer.first * + maxwell3d->regs.index_buffer.FormatSizeInBytes(); runtime.BindIndexBuffer(buffer, new_offset, size); } else { - runtime.BindIndexBuffer(maxwell3d->regs.draw.topology, maxwell3d->regs.index_array.format, - maxwell3d->regs.index_array.first, - maxwell3d->regs.index_array.count, buffer, offset, size); + runtime.BindIndexBuffer(maxwell3d->regs.draw.topology, maxwell3d->regs.index_buffer.format, + maxwell3d->regs.index_buffer.first, + maxwell3d->regs.index_buffer.count, buffer, offset, size); } } @@ -1017,7 +1017,7 @@ void BufferCache<P>::BindHostVertexBuffers() { } flags[Dirty::VertexBuffer0 + index] = false; - const u32 stride = maxwell3d->regs.vertex_array[index].stride; + const u32 stride = maxwell3d->regs.vertex_streams[index].stride; const u32 offset = buffer.Offset(binding.cpu_addr); runtime.BindVertexBuffer(index, buffer, offset, binding.size, stride); } @@ -1157,7 +1157,7 @@ void BufferCache<P>::BindHostGraphicsTextureBuffers(size_t stage) { template <class P> void BufferCache<P>::BindHostTransformFeedbackBuffers() { - if (maxwell3d->regs.tfb_enabled == 0) { + if (maxwell3d->regs.transform_feedback_enabled == 0) { return; } for (u32 index = 0; index < NUM_TRANSFORM_FEEDBACK_BUFFERS; ++index) { @@ -1268,7 +1268,7 @@ template <class P> void BufferCache<P>::UpdateIndexBuffer() { // We have to check for the dirty flags and index count // The index count is currently changed without updating the dirty flags - const auto& index_array = maxwell3d->regs.index_array; + const auto& index_array = maxwell3d->regs.index_buffer; auto& flags = maxwell3d->dirty.flags; if (!flags[Dirty::IndexBuffer] && last_index_count == index_array.count) { return; @@ -1311,10 +1311,10 @@ void BufferCache<P>::UpdateVertexBuffer(u32 index) { if (!maxwell3d->dirty.flags[Dirty::VertexBuffer0 + index]) { return; } - const auto& array = maxwell3d->regs.vertex_array[index]; - const auto& limit = maxwell3d->regs.vertex_array_limit[index]; - const GPUVAddr gpu_addr_begin = array.StartAddress(); - const GPUVAddr gpu_addr_end = limit.LimitAddress() + 1; + const auto& array = maxwell3d->regs.vertex_streams[index]; + const auto& limit = maxwell3d->regs.vertex_stream_limits[index]; + const GPUVAddr gpu_addr_begin = array.Address(); + const GPUVAddr gpu_addr_end = limit.Address() + 1; const std::optional<VAddr> cpu_addr = gpu_memory->GpuToCpuAddress(gpu_addr_begin); u32 address_size = static_cast<u32>( std::min(gpu_addr_end - gpu_addr_begin, static_cast<u64>(std::numeric_limits<u32>::max()))); @@ -1380,7 +1380,7 @@ void BufferCache<P>::UpdateTextureBuffers(size_t stage) { template <class P> void BufferCache<P>::UpdateTransformFeedbackBuffers() { - if (maxwell3d->regs.tfb_enabled == 0) { + if (maxwell3d->regs.transform_feedback_enabled == 0) { return; } for (u32 index = 0; index < NUM_TRANSFORM_FEEDBACK_BUFFERS; ++index) { @@ -1390,11 +1390,11 @@ void BufferCache<P>::UpdateTransformFeedbackBuffers() { template <class P> void BufferCache<P>::UpdateTransformFeedbackBuffer(u32 index) { - const auto& binding = maxwell3d->regs.tfb_bindings[index]; - const GPUVAddr gpu_addr = binding.Address() + binding.buffer_offset; - const u32 size = binding.buffer_size; + const auto& binding = maxwell3d->regs.transform_feedback.buffers[index]; + const GPUVAddr gpu_addr = binding.Address() + binding.start_offset; + const u32 size = binding.size; const std::optional<VAddr> cpu_addr = gpu_memory->GpuToCpuAddress(gpu_addr); - if (binding.buffer_enable == 0 || size == 0 || !cpu_addr) { + if (binding.enable == 0 || size == 0 || !cpu_addr) { transform_feedback_buffers[index] = NULL_BINDING; return; } diff --git a/src/video_core/dirty_flags.cpp b/src/video_core/dirty_flags.cpp index 9dc4341f0..c2ecc12f5 100644 --- a/src/video_core/dirty_flags.cpp +++ b/src/video_core/dirty_flags.cpp @@ -17,21 +17,23 @@ using Tegra::Engines::Maxwell3D; void SetupDirtyVertexBuffers(Maxwell3D::DirtyState::Tables& tables) { static constexpr std::size_t num_array = 3; for (std::size_t i = 0; i < Maxwell3D::Regs::NumVertexArrays; ++i) { - const std::size_t array_offset = OFF(vertex_array) + i * NUM(vertex_array[0]); - const std::size_t limit_offset = OFF(vertex_array_limit) + i * NUM(vertex_array_limit[0]); + const std::size_t array_offset = OFF(vertex_streams) + i * NUM(vertex_streams[0]); + const std::size_t limit_offset = + OFF(vertex_stream_limits) + i * NUM(vertex_stream_limits[0]); FillBlock(tables, array_offset, num_array, VertexBuffer0 + i, VertexBuffers); - FillBlock(tables, limit_offset, NUM(vertex_array_limit), VertexBuffer0 + i, VertexBuffers); + FillBlock(tables, limit_offset, NUM(vertex_stream_limits), VertexBuffer0 + i, + VertexBuffers); } } void SetupIndexBuffer(Maxwell3D::DirtyState::Tables& tables) { - FillBlock(tables[0], OFF(index_array), NUM(index_array), IndexBuffer); + FillBlock(tables[0], OFF(index_buffer), NUM(index_buffer), IndexBuffer); } void SetupDirtyDescriptors(Maxwell3D::DirtyState::Tables& tables) { - FillBlock(tables[0], OFF(tic), NUM(tic), Descriptors); - FillBlock(tables[0], OFF(tsc), NUM(tsc), Descriptors); + FillBlock(tables[0], OFF(tex_header), NUM(tex_header), Descriptors); + FillBlock(tables[0], OFF(tex_sampler), NUM(tex_sampler), Descriptors); } void SetupDirtyRenderTargets(Maxwell3D::DirtyState::Tables& tables) { @@ -42,7 +44,7 @@ void SetupDirtyRenderTargets(Maxwell3D::DirtyState::Tables& tables) { FillBlock(tables[0], begin + rt * num_per_rt, num_per_rt, ColorBuffer0 + rt); } FillBlock(tables[1], begin, num, RenderTargets); - FillBlock(tables[0], OFF(render_area), NUM(render_area), RenderTargets); + FillBlock(tables[0], OFF(surface_clip), NUM(surface_clip), RenderTargets); tables[0][OFF(rt_control)] = RenderTargets; tables[1][OFF(rt_control)] = RenderTargetControl; @@ -52,15 +54,15 @@ void SetupDirtyRenderTargets(Maxwell3D::DirtyState::Tables& tables) { const u8 flag = zeta_flags[i]; auto& table = tables[i]; table[OFF(zeta_enable)] = flag; - table[OFF(zeta_width)] = flag; - table[OFF(zeta_height)] = flag; + table[OFF(zeta_size.width)] = flag; + table[OFF(zeta_size.height)] = flag; FillBlock(table, OFF(zeta), NUM(zeta), flag); } } void SetupDirtyShaders(Maxwell3D::DirtyState::Tables& tables) { - FillBlock(tables[0], OFF(shader_config[0]), - NUM(shader_config[0]) * Maxwell3D::Regs::MaxShaderProgram, Shaders); + FillBlock(tables[0], OFF(pipelines), NUM(pipelines[0]) * Maxwell3D::Regs::MaxShaderProgram, + Shaders); } } // Anonymous namespace diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index 3c6e44a25..89a9d1f5a 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -56,37 +56,37 @@ void Maxwell3D::InitializeRegisterDefaults() { // Doom and Bomberman seems to use the uninitialized registers and just enable blend // so initialize blend registers with sane values - regs.blend.equation_rgb = Regs::Blend::Equation::Add; - regs.blend.factor_source_rgb = Regs::Blend::Factor::One; - regs.blend.factor_dest_rgb = Regs::Blend::Factor::Zero; - regs.blend.equation_a = Regs::Blend::Equation::Add; - regs.blend.factor_source_a = Regs::Blend::Factor::One; - regs.blend.factor_dest_a = Regs::Blend::Factor::Zero; - for (auto& blend : regs.independent_blend) { - blend.equation_rgb = Regs::Blend::Equation::Add; - blend.factor_source_rgb = Regs::Blend::Factor::One; - blend.factor_dest_rgb = Regs::Blend::Factor::Zero; - blend.equation_a = Regs::Blend::Equation::Add; - blend.factor_source_a = Regs::Blend::Factor::One; - blend.factor_dest_a = Regs::Blend::Factor::Zero; - } - regs.stencil_front_op_fail = Regs::StencilOp::Keep; - regs.stencil_front_op_zfail = Regs::StencilOp::Keep; - regs.stencil_front_op_zpass = Regs::StencilOp::Keep; - regs.stencil_front_func_func = Regs::ComparisonOp::Always; + regs.blend.color_op = Regs::Blend::Equation::Add_D3D; + regs.blend.color_source = Regs::Blend::Factor::One_D3D; + regs.blend.color_dest = Regs::Blend::Factor::Zero_D3D; + regs.blend.alpha_op = Regs::Blend::Equation::Add_D3D; + regs.blend.alpha_source = Regs::Blend::Factor::One_D3D; + regs.blend.alpha_dest = Regs::Blend::Factor::Zero_D3D; + for (auto& blend : regs.blend_per_target) { + blend.color_op = Regs::Blend::Equation::Add_D3D; + blend.color_source = Regs::Blend::Factor::One_D3D; + blend.color_dest = Regs::Blend::Factor::Zero_D3D; + blend.alpha_op = Regs::Blend::Equation::Add_D3D; + blend.alpha_source = Regs::Blend::Factor::One_D3D; + blend.alpha_dest = Regs::Blend::Factor::Zero_D3D; + } + regs.stencil_front_op.fail = Regs::StencilOp::Op::Keep_D3D; + regs.stencil_front_op.zfail = Regs::StencilOp::Op::Keep_D3D; + regs.stencil_front_op.zpass = Regs::StencilOp::Op::Keep_D3D; + regs.stencil_front_op.func = Regs::ComparisonOp::Always_GL; regs.stencil_front_func_mask = 0xFFFFFFFF; regs.stencil_front_mask = 0xFFFFFFFF; regs.stencil_two_side_enable = 1; - regs.stencil_back_op_fail = Regs::StencilOp::Keep; - regs.stencil_back_op_zfail = Regs::StencilOp::Keep; - regs.stencil_back_op_zpass = Regs::StencilOp::Keep; - regs.stencil_back_func_func = Regs::ComparisonOp::Always; + regs.stencil_back_op.fail = Regs::StencilOp::Op::Keep_D3D; + regs.stencil_back_op.zfail = Regs::StencilOp::Op::Keep_D3D; + regs.stencil_back_op.zpass = Regs::StencilOp::Op::Keep_D3D; + regs.stencil_back_op.func = Regs::ComparisonOp::Always_GL; regs.stencil_back_func_mask = 0xFFFFFFFF; regs.stencil_back_mask = 0xFFFFFFFF; - regs.depth_test_func = Regs::ComparisonOp::Always; - regs.front_face = Regs::FrontFace::CounterClockWise; - regs.cull_face = Regs::CullFace::Back; + regs.depth_test_func = Regs::ComparisonOp::Always_GL; + regs.gl_front_face = Regs::FrontFace::CounterClockWise; + regs.gl_cull_face = Regs::CullFace::Back; // TODO(Rodrigo): Most games do not set a point size. I think this is a case of a // register carrying a default value. Assume it's OpenGL's default (1). @@ -107,20 +107,20 @@ void Maxwell3D::InitializeRegisterDefaults() { // NVN games expect these values to be enabled at boot regs.rasterize_enable = 1; - regs.rt_separate_frag_data = 1; + regs.color_target_mrt_enable = 1; regs.framebuffer_srgb = 1; regs.line_width_aliased = 1.0f; regs.line_width_smooth = 1.0f; - regs.front_face = Maxwell3D::Regs::FrontFace::ClockWise; + regs.gl_front_face = Maxwell3D::Regs::FrontFace::ClockWise; regs.polygon_mode_back = Maxwell3D::Regs::PolygonMode::Fill; regs.polygon_mode_front = Maxwell3D::Regs::PolygonMode::Fill; shadow_state = regs; - mme_inline[MAXWELL3D_REG_INDEX(draw.vertex_end_gl)] = true; - mme_inline[MAXWELL3D_REG_INDEX(draw.vertex_begin_gl)] = true; + mme_inline[MAXWELL3D_REG_INDEX(draw.end)] = true; + mme_inline[MAXWELL3D_REG_INDEX(draw.begin)] = true; mme_inline[MAXWELL3D_REG_INDEX(vertex_buffer.count)] = true; - mme_inline[MAXWELL3D_REG_INDEX(index_array.count)] = true; + mme_inline[MAXWELL3D_REG_INDEX(index_buffer.count)] = true; } void Maxwell3D::ProcessMacro(u32 method, const u32* base_start, u32 amount, bool is_last_call) { @@ -173,51 +173,56 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume case MAXWELL3D_REG_INDEX(shadow_ram_control): shadow_state.shadow_ram_control = static_cast<Regs::ShadowRamControl>(nonshadow_argument); return; - case MAXWELL3D_REG_INDEX(macros.upload_address): - return macro_engine->ClearCode(regs.macros.upload_address); - case MAXWELL3D_REG_INDEX(macros.data): - return macro_engine->AddCode(regs.macros.upload_address, argument); - case MAXWELL3D_REG_INDEX(macros.bind): + case MAXWELL3D_REG_INDEX(load_mme.instruction_ptr): + return macro_engine->ClearCode(regs.load_mme.instruction_ptr); + case MAXWELL3D_REG_INDEX(load_mme.instruction): + return macro_engine->AddCode(regs.load_mme.instruction_ptr, argument); + case MAXWELL3D_REG_INDEX(load_mme.start_address): return ProcessMacroBind(argument); - case MAXWELL3D_REG_INDEX(firmware[4]): + case MAXWELL3D_REG_INDEX(falcon[4]): return ProcessFirmwareCall4(); - case MAXWELL3D_REG_INDEX(const_buffer.cb_data): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 1: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 2: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 3: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 4: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 5: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 6: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 7: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 8: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 9: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 10: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 11: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 12: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 13: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 14: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 15: + case MAXWELL3D_REG_INDEX(const_buffer.buffer): + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 1: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 2: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 3: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 4: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 5: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 6: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 7: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 8: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 9: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 10: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 11: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 12: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 13: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 14: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 15: return ProcessCBData(argument); - case MAXWELL3D_REG_INDEX(cb_bind[0]): + case MAXWELL3D_REG_INDEX(bind_groups[0].raw_config): return ProcessCBBind(0); - case MAXWELL3D_REG_INDEX(cb_bind[1]): + case MAXWELL3D_REG_INDEX(bind_groups[1].raw_config): return ProcessCBBind(1); - case MAXWELL3D_REG_INDEX(cb_bind[2]): + case MAXWELL3D_REG_INDEX(bind_groups[2].raw_config): return ProcessCBBind(2); - case MAXWELL3D_REG_INDEX(cb_bind[3]): + case MAXWELL3D_REG_INDEX(bind_groups[3].raw_config): return ProcessCBBind(3); - case MAXWELL3D_REG_INDEX(cb_bind[4]): + case MAXWELL3D_REG_INDEX(bind_groups[4].raw_config): return ProcessCBBind(4); - case MAXWELL3D_REG_INDEX(draw.vertex_end_gl): + case MAXWELL3D_REG_INDEX(draw.end): return DrawArrays(); - case MAXWELL3D_REG_INDEX(small_index): - regs.index_array.count = regs.small_index.count; - regs.index_array.first = regs.small_index.first; + case MAXWELL3D_REG_INDEX(index_buffer32_first): + regs.index_buffer.count = regs.index_buffer32_first.count; + regs.index_buffer.first = regs.index_buffer32_first.first; dirty.flags[VideoCommon::Dirty::IndexBuffer] = true; return DrawArrays(); - case MAXWELL3D_REG_INDEX(small_index_2): - regs.index_array.count = regs.small_index_2.count; - regs.index_array.first = regs.small_index_2.first; + case MAXWELL3D_REG_INDEX(index_buffer16_first): + regs.index_buffer.count = regs.index_buffer16_first.count; + regs.index_buffer.first = regs.index_buffer16_first.first; + dirty.flags[VideoCommon::Dirty::IndexBuffer] = true; + return DrawArrays(); + case MAXWELL3D_REG_INDEX(index_buffer8_first): + regs.index_buffer.count = regs.index_buffer8_first.count; + regs.index_buffer.first = regs.index_buffer8_first.first; dirty.flags[VideoCommon::Dirty::IndexBuffer] = true; // a macro calls this one over and over, should it increase instancing? // Used by Hades and likely other Vulkan games. @@ -225,28 +230,24 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume case MAXWELL3D_REG_INDEX(topology_override): use_topology_override = true; return; - case MAXWELL3D_REG_INDEX(clear_buffers): + case MAXWELL3D_REG_INDEX(clear_surface): return ProcessClearBuffers(); - case MAXWELL3D_REG_INDEX(query.query_get): + case MAXWELL3D_REG_INDEX(report_semaphore.query): return ProcessQueryGet(); - case MAXWELL3D_REG_INDEX(condition.mode): + case MAXWELL3D_REG_INDEX(render_enable.mode): return ProcessQueryCondition(); - case MAXWELL3D_REG_INDEX(counter_reset): + case MAXWELL3D_REG_INDEX(clear_report_value): return ProcessCounterReset(); case MAXWELL3D_REG_INDEX(sync_info): return ProcessSyncPoint(); - case MAXWELL3D_REG_INDEX(exec_upload): - return upload_state.ProcessExec(regs.exec_upload.linear != 0); - case MAXWELL3D_REG_INDEX(data_upload): + case MAXWELL3D_REG_INDEX(launch_dma): + return upload_state.ProcessExec(regs.launch_dma.memory_layout.Value() == + Regs::LaunchDMA::Layout::Pitch); + case MAXWELL3D_REG_INDEX(inline_data): upload_state.ProcessData(argument, is_last_call); return; case MAXWELL3D_REG_INDEX(fragment_barrier): return rasterizer->FragmentBarrier(); - case MAXWELL3D_REG_INDEX(invalidate_texture_data_cache): - rasterizer->InvalidateGPUCache(); - return rasterizer->WaitForIdle(); - case MAXWELL3D_REG_INDEX(tiled_cache_barrier): - return rasterizer->TiledCacheBarrier(); } } @@ -296,25 +297,25 @@ void Maxwell3D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, return; } switch (method) { - case MAXWELL3D_REG_INDEX(const_buffer.cb_data): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 1: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 2: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 3: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 4: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 5: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 6: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 7: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 8: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 9: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 10: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 11: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 12: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 13: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 14: - case MAXWELL3D_REG_INDEX(const_buffer.cb_data) + 15: + case MAXWELL3D_REG_INDEX(const_buffer.buffer): + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 1: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 2: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 3: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 4: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 5: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 6: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 7: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 8: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 9: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 10: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 11: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 12: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 13: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 14: + case MAXWELL3D_REG_INDEX(const_buffer.buffer) + 15: ProcessCBMultiData(base_start, amount); break; - case MAXWELL3D_REG_INDEX(data_upload): + case MAXWELL3D_REG_INDEX(inline_data): upload_state.ProcessData(base_start, static_cast<size_t>(amount)); return; default: @@ -353,14 +354,15 @@ void Maxwell3D::CallMethodFromMME(u32 method, u32 method_argument) { if (mme_inline[method]) { regs.reg_array[method] = method_argument; if (method == MAXWELL3D_REG_INDEX(vertex_buffer.count) || - method == MAXWELL3D_REG_INDEX(index_array.count)) { + method == MAXWELL3D_REG_INDEX(index_buffer.count)) { const MMEDrawMode expected_mode = method == MAXWELL3D_REG_INDEX(vertex_buffer.count) ? MMEDrawMode::Array : MMEDrawMode::Indexed; StepInstance(expected_mode, method_argument); - } else if (method == MAXWELL3D_REG_INDEX(draw.vertex_begin_gl)) { + } else if (method == MAXWELL3D_REG_INDEX(draw.begin)) { mme_draw.instance_mode = - (regs.draw.instance_next != 0) || (regs.draw.instance_cont != 0); + (regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Subsequent) || + (regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Unchanged); mme_draw.gl_begin_consume = true; } else { mme_draw.gl_end_count++; @@ -405,11 +407,12 @@ void Maxwell3D::ProcessTopologyOverride() { void Maxwell3D::FlushMMEInlineDraw() { LOG_TRACE(HW_GPU, "called, topology={}, count={}", regs.draw.topology.Value(), regs.vertex_buffer.count); - ASSERT_MSG(!(regs.index_array.count && regs.vertex_buffer.count), "Both indexed and direct?"); + ASSERT_MSG(!(regs.index_buffer.count && regs.vertex_buffer.count), "Both indexed and direct?"); ASSERT(mme_draw.instance_count == mme_draw.gl_end_count); // Both instance configuration registers can not be set at the same time. - ASSERT_MSG(!regs.draw.instance_next || !regs.draw.instance_cont, + ASSERT_MSG(regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::First || + regs.draw.instance_id != Maxwell3D::Regs::Draw::InstanceId::Unchanged, "Illegal combination of instancing parameters"); ProcessTopologyOverride(); @@ -424,7 +427,7 @@ void Maxwell3D::FlushMMEInlineDraw() { // it's possible that it is incorrect and that there is some other register used to specify the // drawing mode. if (is_indexed) { - regs.index_array.count = 0; + regs.index_buffer.count = 0; } else { regs.vertex_buffer.count = 0; } @@ -437,11 +440,11 @@ void Maxwell3D::FlushMMEInlineDraw() { } void Maxwell3D::ProcessMacroUpload(u32 data) { - macro_engine->AddCode(regs.macros.upload_address++, data); + macro_engine->AddCode(regs.load_mme.instruction_ptr++, data); } void Maxwell3D::ProcessMacroBind(u32 data) { - macro_positions[regs.macros.entry++] = data; + macro_positions[regs.load_mme.start_address_ptr++] = data; } void Maxwell3D::ProcessFirmwareCall4() { @@ -449,11 +452,11 @@ void Maxwell3D::ProcessFirmwareCall4() { // Firmware call 4 is a blob that changes some registers depending on its parameters. // These registers don't affect emulation and so are stubbed by setting 0xd00 to 1. - regs.reg_array[0xd00] = 1; + regs.shadow_scratch[0] = 1; } void Maxwell3D::StampQueryResult(u64 payload, bool long_query) { - const GPUVAddr sequence_address{regs.query.QueryAddress()}; + const GPUVAddr sequence_address{regs.report_semaphore.Address()}; if (long_query) { memory_manager.Write<u64>(sequence_address + sizeof(u64), system.GPU().GetTicks()); memory_manager.Write<u64>(sequence_address, payload); @@ -464,15 +467,17 @@ void Maxwell3D::StampQueryResult(u64 payload, bool long_query) { void Maxwell3D::ProcessQueryGet() { // TODO(Subv): Support the other query units. - if (regs.query.query_get.unit != Regs::QueryUnit::Crop) { - LOG_DEBUG(HW_GPU, "Units other than CROP are unimplemented"); + if (regs.report_semaphore.query.location != Regs::ReportSemaphore::Location::All) { + LOG_DEBUG(HW_GPU, "Locations other than ALL are unimplemented"); } - switch (regs.query.query_get.operation) { - case Regs::QueryOperation::Release: - if (regs.query.query_get.fence == 1 || regs.query.query_get.short_query != 0) { - const GPUVAddr sequence_address{regs.query.QueryAddress()}; - const u32 payload = regs.query.query_sequence; + switch (regs.report_semaphore.query.operation) { + case Regs::ReportSemaphore::Operation::Release: + if (regs.report_semaphore.query.release == + Regs::ReportSemaphore::Release::AfterAllPreceedingWrites || + regs.report_semaphore.query.short_query != 0) { + const GPUVAddr sequence_address{regs.report_semaphore.Address()}; + const u32 payload = regs.report_semaphore.payload; std::function<void()> operation([this, sequence_address, payload] { memory_manager.Write<u32>(sequence_address, payload); }); @@ -482,8 +487,8 @@ void Maxwell3D::ProcessQueryGet() { u64_le value; u64_le timestamp; }; - const GPUVAddr sequence_address{regs.query.QueryAddress()}; - const u32 payload = regs.query.query_sequence; + const GPUVAddr sequence_address{regs.report_semaphore.Address()}; + const u32 payload = regs.report_semaphore.payload; std::function<void()> operation([this, sequence_address, payload] { memory_manager.Write<u64>(sequence_address + sizeof(u64), system.GPU().GetTicks()); memory_manager.Write<u64>(sequence_address, payload); @@ -491,19 +496,19 @@ void Maxwell3D::ProcessQueryGet() { rasterizer->SyncOperation(std::move(operation)); } break; - case Regs::QueryOperation::Acquire: + case Regs::ReportSemaphore::Operation::Acquire: // TODO(Blinkhawk): Under this operation, the GPU waits for the CPU to write a value that // matches the current payload. UNIMPLEMENTED_MSG("Unimplemented query operation ACQUIRE"); break; - case Regs::QueryOperation::Counter: + case Regs::ReportSemaphore::Operation::ReportOnly: if (const std::optional<u64> result = GetQueryResult()) { // If the query returns an empty optional it means it's cached and deferred. // In this case we have a non-empty result, so we stamp it immediately. - StampQueryResult(*result, regs.query.query_get.short_query == 0); + StampQueryResult(*result, regs.report_semaphore.query.short_query == 0); } break; - case Regs::QueryOperation::Trap: + case Regs::ReportSemaphore::Operation::Trap: UNIMPLEMENTED_MSG("Unimplemented query operation TRAP"); break; default: @@ -513,31 +518,31 @@ void Maxwell3D::ProcessQueryGet() { } void Maxwell3D::ProcessQueryCondition() { - const GPUVAddr condition_address{regs.condition.Address()}; - switch (regs.condition.mode) { - case Regs::ConditionMode::Always: { + const GPUVAddr condition_address{regs.render_enable.Address()}; + switch (regs.render_enable.mode) { + case Regs::RenderEnable::Mode::True: { execute_on = true; break; } - case Regs::ConditionMode::Never: { + case Regs::RenderEnable::Mode::False: { execute_on = false; break; } - case Regs::ConditionMode::ResNonZero: { - Regs::QueryCompare cmp; + case Regs::RenderEnable::Mode::Conditional: { + Regs::ReportSemaphore::Compare cmp; memory_manager.ReadBlock(condition_address, &cmp, sizeof(cmp)); execute_on = cmp.initial_sequence != 0U && cmp.initial_mode != 0U; break; } - case Regs::ConditionMode::Equal: { - Regs::QueryCompare cmp; + case Regs::RenderEnable::Mode::IfEqual: { + Regs::ReportSemaphore::Compare cmp; memory_manager.ReadBlock(condition_address, &cmp, sizeof(cmp)); execute_on = cmp.initial_sequence == cmp.current_sequence && cmp.initial_mode == cmp.current_mode; break; } - case Regs::ConditionMode::NotEqual: { - Regs::QueryCompare cmp; + case Regs::RenderEnable::Mode::IfNotEqual: { + Regs::ReportSemaphore::Compare cmp; memory_manager.ReadBlock(condition_address, &cmp, sizeof(cmp)); execute_on = cmp.initial_sequence != cmp.current_sequence || cmp.initial_mode != cmp.current_mode; @@ -552,21 +557,21 @@ void Maxwell3D::ProcessQueryCondition() { } void Maxwell3D::ProcessCounterReset() { - switch (regs.counter_reset) { - case Regs::CounterReset::SampleCnt: + switch (regs.clear_report_value) { + case Regs::ClearReport::ZPassPixelCount: rasterizer->ResetCounter(QueryType::SamplesPassed); break; default: - LOG_DEBUG(Render_OpenGL, "Unimplemented counter reset={}", regs.counter_reset); + LOG_DEBUG(Render_OpenGL, "Unimplemented counter reset={}", regs.clear_report_value); break; } } void Maxwell3D::ProcessSyncPoint() { const u32 sync_point = regs.sync_info.sync_point.Value(); - const u32 increment = regs.sync_info.increment.Value(); - [[maybe_unused]] const u32 cache_flush = regs.sync_info.unknown.Value(); - if (increment) { + const auto condition = regs.sync_info.condition.Value(); + [[maybe_unused]] const u32 cache_flush = regs.sync_info.clean_l2.Value(); + if (condition == Regs::SyncInfo::Condition::RopWritesDone) { rasterizer->SignalSyncPoint(sync_point); } } @@ -574,23 +579,24 @@ void Maxwell3D::ProcessSyncPoint() { void Maxwell3D::DrawArrays() { LOG_TRACE(HW_GPU, "called, topology={}, count={}", regs.draw.topology.Value(), regs.vertex_buffer.count); - ASSERT_MSG(!(regs.index_array.count && regs.vertex_buffer.count), "Both indexed and direct?"); + ASSERT_MSG(!(regs.index_buffer.count && regs.vertex_buffer.count), "Both indexed and direct?"); // Both instance configuration registers can not be set at the same time. - ASSERT_MSG(!regs.draw.instance_next || !regs.draw.instance_cont, + ASSERT_MSG(regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::First || + regs.draw.instance_id != Maxwell3D::Regs::Draw::InstanceId::Unchanged, "Illegal combination of instancing parameters"); ProcessTopologyOverride(); - if (regs.draw.instance_next) { + if (regs.draw.instance_id == Maxwell3D::Regs::Draw::InstanceId::Subsequent) { // Increment the current instance *before* drawing. - state.current_instance += 1; - } else if (!regs.draw.instance_cont) { + state.current_instance++; + } else if (regs.draw.instance_id != Maxwell3D::Regs::Draw::InstanceId::Unchanged) { // Reset the current instance to 0. state.current_instance = 0; } - const bool is_indexed{regs.index_array.count && !regs.vertex_buffer.count}; + const bool is_indexed{regs.index_buffer.count && !regs.vertex_buffer.count}; if (ShouldExecute()) { rasterizer->Draw(is_indexed, false); } @@ -600,60 +606,60 @@ void Maxwell3D::DrawArrays() { // it's possible that it is incorrect and that there is some other register used to specify the // drawing mode. if (is_indexed) { - regs.index_array.count = 0; + regs.index_buffer.count = 0; } else { regs.vertex_buffer.count = 0; } } std::optional<u64> Maxwell3D::GetQueryResult() { - switch (regs.query.query_get.select) { - case Regs::QuerySelect::Payload: - return regs.query.query_sequence; - case Regs::QuerySelect::SamplesPassed: + switch (regs.report_semaphore.query.report) { + case Regs::ReportSemaphore::Report::Payload: + return regs.report_semaphore.payload; + case Regs::ReportSemaphore::Report::ZPassPixelCount64: // Deferred. - rasterizer->Query(regs.query.QueryAddress(), QueryType::SamplesPassed, + rasterizer->Query(regs.report_semaphore.Address(), QueryType::SamplesPassed, system.GPU().GetTicks()); return std::nullopt; default: - LOG_DEBUG(HW_GPU, "Unimplemented query select type {}", - regs.query.query_get.select.Value()); + LOG_DEBUG(HW_GPU, "Unimplemented query report type {}", + regs.report_semaphore.query.report.Value()); return 1; } } void Maxwell3D::ProcessCBBind(size_t stage_index) { // Bind the buffer currently in CB_ADDRESS to the specified index in the desired shader stage. - const auto& bind_data = regs.cb_bind[stage_index]; - auto& buffer = state.shader_stages[stage_index].const_buffers[bind_data.index]; + const auto& bind_data = regs.bind_groups[stage_index]; + auto& buffer = state.shader_stages[stage_index].const_buffers[bind_data.shader_slot]; buffer.enabled = bind_data.valid.Value() != 0; - buffer.address = regs.const_buffer.BufferAddress(); - buffer.size = regs.const_buffer.cb_size; + buffer.address = regs.const_buffer.Address(); + buffer.size = regs.const_buffer.size; const bool is_enabled = bind_data.valid.Value() != 0; if (!is_enabled) { - rasterizer->DisableGraphicsUniformBuffer(stage_index, bind_data.index); + rasterizer->DisableGraphicsUniformBuffer(stage_index, bind_data.shader_slot); return; } - const GPUVAddr gpu_addr = regs.const_buffer.BufferAddress(); - const u32 size = regs.const_buffer.cb_size; - rasterizer->BindGraphicsUniformBuffer(stage_index, bind_data.index, gpu_addr, size); + const GPUVAddr gpu_addr = regs.const_buffer.Address(); + const u32 size = regs.const_buffer.size; + rasterizer->BindGraphicsUniformBuffer(stage_index, bind_data.shader_slot, gpu_addr, size); } void Maxwell3D::ProcessCBMultiData(const u32* start_base, u32 amount) { // Write the input value to the current const buffer at the current position. - const GPUVAddr buffer_address = regs.const_buffer.BufferAddress(); + const GPUVAddr buffer_address = regs.const_buffer.Address(); ASSERT(buffer_address != 0); // Don't allow writing past the end of the buffer. - ASSERT(regs.const_buffer.cb_pos <= regs.const_buffer.cb_size); + ASSERT(regs.const_buffer.offset <= regs.const_buffer.size); - const GPUVAddr address{buffer_address + regs.const_buffer.cb_pos}; + const GPUVAddr address{buffer_address + regs.const_buffer.offset}; const size_t copy_size = amount * sizeof(u32); memory_manager.WriteBlock(address, start_base, copy_size); // Increment the current buffer position. - regs.const_buffer.cb_pos += static_cast<u32>(copy_size); + regs.const_buffer.offset += static_cast<u32>(copy_size); } void Maxwell3D::ProcessCBData(u32 value) { @@ -661,7 +667,8 @@ void Maxwell3D::ProcessCBData(u32 value) { } Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const { - const GPUVAddr tic_address_gpu{regs.tic.Address() + tic_index * sizeof(Texture::TICEntry)}; + const GPUVAddr tic_address_gpu{regs.tex_header.Address() + + tic_index * sizeof(Texture::TICEntry)}; Texture::TICEntry tic_entry; memory_manager.ReadBlockUnsafe(tic_address_gpu, &tic_entry, sizeof(Texture::TICEntry)); @@ -670,7 +677,8 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const { } Texture::TSCEntry Maxwell3D::GetTSCEntry(u32 tsc_index) const { - const GPUVAddr tsc_address_gpu{regs.tsc.Address() + tsc_index * sizeof(Texture::TSCEntry)}; + const GPUVAddr tsc_address_gpu{regs.tex_sampler.Address() + + tsc_index * sizeof(Texture::TSCEntry)}; Texture::TSCEntry tsc_entry; memory_manager.ReadBlockUnsafe(tsc_address_gpu, &tsc_entry, sizeof(Texture::TSCEntry)); diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index 5f9eb208c..12dbd9cc4 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -39,12 +39,15 @@ namespace Tegra::Engines { /** * This Engine is known as GF100_3D. Documentation can be found in: + * https://github.com/NVIDIA/open-gpu-doc/blob/master/classes/3d/clb197.h * https://github.com/envytools/envytools/blob/master/rnndb/graph/gf100_3d.xml * https://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/drivers/nouveau/nvc0/nvc0_3d.xml.h + * + * Note: nVidia have confirmed that their open docs have had parts redacted, so this list is + * currently incomplete, and the gaps are still worth exploring. */ -#define MAXWELL3D_REG_INDEX(field_name) \ - (offsetof(Tegra::Engines::Maxwell3D::Regs, field_name) / sizeof(u32)) +#define MAXWELL3D_REG_INDEX(field_name) (offsetof(Maxwell3D::Regs, field_name) / sizeof(u32)) class Maxwell3D final : public EngineInterface { public: @@ -55,7 +58,6 @@ public: void BindRasterizer(VideoCore::RasterizerInterface* rasterizer); /// Register structure of the Maxwell3D engine. - /// TODO(Subv): This structure will need to be made bigger as more registers are discovered. struct Regs { static constexpr std::size_t NUM_REGS = 0xE00; @@ -74,90 +76,515 @@ public: static constexpr std::size_t MaxConstBuffers = 18; static constexpr std::size_t MaxConstBufferSize = 0x10000; - enum class QueryOperation : u32 { - Release = 0, - Acquire = 1, - Counter = 2, - Trap = 3, + struct ID { + union { + BitField<0, 16, u32> cls; + BitField<16, 5, u32> engine; + }; }; - enum class QueryUnit : u32 { - VFetch = 1, - VP = 2, - Rast = 4, - StrmOut = 5, - GP = 6, - ZCull = 7, - Prop = 10, - Crop = 15, + struct LoadMME { + u32 instruction_ptr; + u32 instruction; + u32 start_address_ptr; + u32 start_address; }; - enum class QuerySelect : u32 { - Payload = 0, - TimeElapsed = 2, - TransformFeedbackPrimitivesGenerated = 11, - PrimitivesGenerated = 18, - SamplesPassed = 21, - TransformFeedbackUnknown = 26, + struct Notify { + u32 address_high; + u32 address_low; + u32 type; + + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } }; - struct QueryCompare { - u32 initial_sequence; - u32 initial_mode; - u32 unknown1; - u32 unknown2; - u32 current_sequence; - u32 current_mode; + struct PeerSemaphore { + u32 address_high; + u32 address_low; + + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } }; - enum class QuerySyncCondition : u32 { - NotEqual = 0, - GreaterThan = 1, + struct GlobalRender { + enum class Mode : u32 { + False = 0, + True = 1, + Conditional = 2, + IfEqual = 3, + IfNotEqual = 4, + }; + u32 offset_high; + u32 offset_low; + Mode mode; + + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(offset_high) << 32) | + offset_low); + } }; - enum class ConditionMode : u32 { - Never = 0, - Always = 1, - ResNonZero = 2, - Equal = 3, - NotEqual = 4, + enum class ReductionOp : u32 { + RedAdd = 0, + RedMin = 1, + RedMax = 2, + RedInc = 3, + RedDec = 4, + RedAnd = 5, + RedOr = 6, + RedXor = 7, }; - enum class ShaderProgram : u32 { - VertexA = 0, - VertexB = 1, - TesselationControl = 2, - TesselationEval = 3, - Geometry = 4, - Fragment = 5, + struct LaunchDMA { + enum class Layout : u32 { + Blocklinear = 0, + Pitch = 1, + }; + + enum class CompletionType : u32 { + FlushDisable = 0, + FlushOnly = 1, + Release = 2, + }; + + union { + BitField<0, 1, Layout> memory_layout; + BitField<4, 2, CompletionType> completion_type; + BitField<8, 2, u32> interrupt_type; + BitField<12, 1, u32> sem_size; + BitField<1, 1, u32> reduction_enable; + BitField<13, 3, ReductionOp> reduction_op; + BitField<2, 2, u32> reduction_format; + BitField<6, 1, u32> sysmembar_disable; + }; + }; + + struct I2M { + u32 address_high; + u32 address_low; + u32 payload; + INSERT_PADDING_BYTES_NOINIT(0x8); + u32 nop0; + u32 nop1; + u32 nop2; + u32 nop3; + }; + + struct OpportunisticEarlyZ { + BitField<0, 5, u32> threshold; + + u32 Threshold() const { + switch (threshold) { + case 0x0: + return 0; + case 0x1F: + return 0x1F; + default: + // Thresholds begin at 0x10 (1 << 4) + // Threshold is in the range 0x1 to 0x13 + return 1 << (4 + threshold.Value() - 1); + } + } + }; + + struct GeometryShaderDmFifo { + union { + BitField<0, 13, u32> raster_on; + BitField<16, 13, u32> raster_off; + BitField<31, 1, u32> spill_enabled; + }; + }; + + struct L2CacheControl { + enum class EvictPolicy : u32 { + First = 0, + Normal = 1, + Last = 2, + }; + + union { + BitField<4, 2, EvictPolicy> policy; + }; + }; + + struct InvalidateShaderCache { + union { + BitField<0, 1, u32> instruction; + BitField<4, 1, u32> data; + BitField<12, 1, u32> constant; + BitField<1, 1, u32> locks; + BitField<2, 1, u32> flush_data; + }; + }; + + struct SyncInfo { + enum class Condition : u32 { + StreamOutWritesDone = 0, + RopWritesDone = 1, + }; + + union { + BitField<0, 16, u32> sync_point; + BitField<16, 1, u32> clean_l2; + BitField<20, 1, Condition> condition; + }; + }; + + struct SurfaceClipBlockId { + union { + BitField<0, 4, u32> block_width; + BitField<4, 4, u32> block_height; + BitField<8, 4, u32> block_depth; + }; + }; + + struct DecompressSurface { + union { + BitField<0, 3, u32> mrt_select; + BitField<4, 16, u32> rt_array_index; + }; + }; + + struct ZCullRopBypass { + union { + BitField<0, 1, u32> enable; + BitField<4, 1, u32> no_stall; + BitField<8, 1, u32> cull_everything; + BitField<12, 4, u32> threshold; + }; + }; + + struct ZCullSubregion { + union { + BitField<0, 1, u32> enable; + BitField<4, 24, u32> normalized_aliquots; + }; + }; + + struct RasterBoundingBox { + enum class Mode : u32 { + BoundingBox = 0, + FullViewport = 1, + }; + + union { + BitField<0, 1, Mode> mode; + BitField<4, 8, u32> pad; + }; + }; + + struct IteratedBlendOptimization { + enum class Noop : u32 { + Never = 0, + SourceRGBA0000 = 1, + SourceAlpha = 2, + SourceRGBA0001 = 3, + }; + + union { + BitField<0, 1, Noop> noop; + }; + }; + + struct ZCullSubregionAllocation { + enum class Format : u32 { + Z_16x16x2_4x4 = 0, + ZS_16x16_4x4 = 1, + Z_16x16_4x2 = 2, + Z_16x16_2x4 = 3, + Z_16x8_4x4 = 4, + Z_8x8_4x2 = 5, + Z_8x8_2x4 = 6, + Z_16x16_4x8 = 7, + Z_4x8_2x2 = 8, + ZS_16x8_4x2 = 9, + ZS_16x8_2x4 = 10, + ZS_8x8_2x2 = 11, + Z_4x8_1x1 = 12, + None = 15, + }; + + union { + BitField<0, 8, u32> id; + BitField<8, 16, u32> aliquots; + BitField<24, 4, Format> format; + }; + }; + + enum class ZCullSubregionAlgorithm : u32 { + Static = 0, + Adaptive = 1, + }; + + struct PixelShaderOutputSampleMaskUsage { + union { + BitField<0, 1, u32> enable; + BitField<1, 1, u32> qualify_by_aa; + }; + }; + + struct L1Configuration { + enum class AddressableMemory : u32 { + Size16Kb = 0, + Size48Kb = 3, + }; + union { + BitField<0, 3, AddressableMemory> direct_addressable_memory; + }; + }; + + struct SPAVersion { + union { + BitField<0, 8, u32> minor; + BitField<8, 8, u32> major; + }; + }; + + struct SnapGrid { + enum class Location : u32 { + Pixel2x2 = 1, + Pixel4x4 = 2, + Pixel8x8 = 3, + Pixel16x16 = 4, + Pixel32x32 = 5, + Pixel64x64 = 6, + Pixel128x128 = 7, + Pixel256x256 = 8, + }; + + enum class Mode : u32 { + RTNE = 0, + Tesla = 1, + }; + + struct { + union { + BitField<0, 4, Location> location; + BitField<8, 1, Mode> rounding_mode; + }; + } line; + + struct { + union { + BitField<0, 4, Location> location; + BitField<8, 1, Mode> rounding_mode; + }; + } non_line; + }; + + struct Tessellation { + enum class DomainType : u32 { + Isolines = 0, + Triangles = 1, + Quads = 2, + }; + + enum class Spacing : u32 { + Integer = 0, + FractionalOdd = 1, + FractionalEven = 2, + }; + + enum class OutputPrimitves : u32 { + Points = 0, + Lines = 1, + Triangles_CW = 2, + Triangles_CCW = 3, + }; + + struct Parameters { + union { + BitField<0, 2, DomainType> domain_type; + BitField<4, 2, Spacing> spacing; + BitField<8, 2, OutputPrimitves> output_primitives; + }; + } params; + + struct LOD { + std::array<f32, 4> outer; + std::array<f32, 2> inner; + } lod; + + std::array<u32, 9> reserved; + }; + + struct SubTilingPerf { + struct { + union { + BitField<0, 8, u32> spm_triangle_register_file_per; + BitField<8, 8, u32> spm_pixel_output_buffer_per; + BitField<16, 8, u32> spm_triangle_ram_per; + BitField<24, 8, u32> max_quads_per; + }; + } knob_a; + + struct { + union { + BitField<0, 8, u32> max_primitives_per; + }; + } knob_b; + + u32 knob_c; + }; + + struct ZCullSubregionReport { + enum class ReportType : u32 { + DepthTest = 0, + DepthTestNoAccept = 1, + DepthTestLateZ = 2, + StencilTest = 3, + }; + + union { + BitField<0, 1, u32> enabled; + BitField<4, 8, u32> subregion_id; + } to_report; + + union { + BitField<0, 1, u32> enabled; + BitField<4, 3, ReportType> type; + } report_type; + }; + + struct BalancedPrimitiveWorkload { + union { + BitField<0, 1, u32> unpartitioned_mode; + BitField<4, 1, u32> timesliced_mode; + }; + }; + + struct TransformFeedback { + struct Buffer { + u32 enable; + u32 address_high; + u32 address_low; + s32 size; + s32 start_offset; + INSERT_PADDING_BYTES_NOINIT(0xC); + + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; + static_assert(sizeof(Buffer) == 0x20); + + struct Control { + u32 stream; + u32 varying_count; + u32 stride; + INSERT_PADDING_BYTES_NOINIT(0x4); + }; + static_assert(sizeof(Control) == 0x10); + + std::array<TransformFeedback::Buffer, NumTransformFeedbackBuffers> buffers; + + INSERT_PADDING_BYTES_NOINIT(0x300); + + std::array<TransformFeedback::Control, NumTransformFeedbackBuffers> controls; + }; + + struct HybridAntiAliasControl { + enum class Centroid : u32 { + PerFragment = 0, + PerPass = 1, + }; + union { + BitField<0, 4, u32> passes; + BitField<4, 1, Centroid> centroid; + BitField<5, 1, u32> passes_extended; + }; + }; + + struct ShaderLocalMemory { + u32 base_address; + INSERT_PADDING_BYTES_NOINIT(0x10); + u32 address_high; + u32 address_low; + u32 size_high; + u32 size_low; + u32 default_size_per_warp; + + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + + u64 Size() const { + return (static_cast<u64>(size_high) << 32) | size_low; + } + }; + + struct ZCullRegion { + u32 width; + u32 height; + u32 depth; + u32 offset; + INSERT_PADDING_BYTES_NOINIT(0xC); + u32 fetch_streams_once; + union { + BitField<0, 16, u32> start_aliquot; + BitField<16, 16, u32> aliquot_count; + } location; + u32 aliquots_per_layer; + u32 storage_address_high; + u32 storage_address_low; + u32 storage_limit_address_high; + u32 storage_limit_address_low; + + GPUVAddr StorageAddress() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(storage_address_high) << 32) | + storage_address_low); + } + GPUVAddr StorageLimitAddress() const { + return static_cast<GPUVAddr>( + (static_cast<GPUVAddr>(storage_limit_address_high) << 32) | + storage_limit_address_low); + } + }; + + struct ZetaReadOnly { + union { + BitField<0, 1, u32> enable_z; + BitField<4, 1, u32> enable_stencil; + }; }; struct VertexAttribute { enum class Size : u32 { Invalid = 0x0, - Size_32_32_32_32 = 0x01, - Size_32_32_32 = 0x02, - Size_16_16_16_16 = 0x03, - Size_32_32 = 0x04, - Size_16_16_16 = 0x05, - Size_8_8_8_8 = 0x0a, - Size_16_16 = 0x0f, - Size_32 = 0x12, - Size_8_8_8 = 0x13, - Size_8_8 = 0x18, - Size_16 = 0x1b, - Size_8 = 0x1d, - Size_10_10_10_2 = 0x30, - Size_11_11_10 = 0x31, + Size_R32_G32_B32_A32 = 0x01, + Size_R32_G32_B32 = 0x02, + Size_R16_G16_B16_A16 = 0x03, + Size_R32_G32 = 0x04, + Size_R16_G16_B16 = 0x05, + Size_R8_G8_B8_A8 = 0x0A, + Size_R16_G16 = 0x0F, + Size_R32 = 0x12, + Size_R8_G8_B8 = 0x13, + Size_R8_G8 = 0x18, + Size_R16 = 0x1B, + Size_R8 = 0x1D, + Size_A2_B10_G10_R10 = 0x30, + Size_B10_G11_R11 = 0x31, + Size_G8_R8 = 0x32, + Size_X8_B8_G8_R8 = 0x33, + Size_A8 = 0x34, }; enum class Type : u32 { - SignedNorm = 1, - UnsignedNorm = 2, - SignedInt = 3, - UnsignedInt = 4, - UnsignedScaled = 5, - SignedScaled = 6, + UnusedEnumDoNotUseBecauseItWillGoAway = 0, + SNorm = 1, + UNorm = 2, + SInt = 3, + UInt = 4, + UScaled = 5, + SScaled = 6, Float = 7, }; @@ -173,33 +600,36 @@ public: u32 ComponentCount() const { switch (size) { - case Size::Size_32_32_32_32: + case Size::Size_R32_G32_B32_A32: return 4; - case Size::Size_32_32_32: + case Size::Size_R32_G32_B32: return 3; - case Size::Size_16_16_16_16: + case Size::Size_R16_G16_B16_A16: return 4; - case Size::Size_32_32: + case Size::Size_R32_G32: return 2; - case Size::Size_16_16_16: + case Size::Size_R16_G16_B16: return 3; - case Size::Size_8_8_8_8: + case Size::Size_R8_G8_B8_A8: + case Size::Size_X8_B8_G8_R8: return 4; - case Size::Size_16_16: + case Size::Size_R16_G16: return 2; - case Size::Size_32: + case Size::Size_R32: return 1; - case Size::Size_8_8_8: + case Size::Size_R8_G8_B8: return 3; - case Size::Size_8_8: + case Size::Size_R8_G8: + case Size::Size_G8_R8: return 2; - case Size::Size_16: + case Size::Size_R16: return 1; - case Size::Size_8: + case Size::Size_R8: + case Size::Size_A8: return 1; - case Size::Size_10_10_10_2: + case Size::Size_A2_B10_G10_R10: return 4; - case Size::Size_11_11_10: + case Size::Size_B10_G11_R11: return 3; default: ASSERT(false); @@ -209,33 +639,36 @@ public: u32 SizeInBytes() const { switch (size) { - case Size::Size_32_32_32_32: + case Size::Size_R32_G32_B32_A32: return 16; - case Size::Size_32_32_32: + case Size::Size_R32_G32_B32: return 12; - case Size::Size_16_16_16_16: + case Size::Size_R16_G16_B16_A16: return 8; - case Size::Size_32_32: + case Size::Size_R32_G32: return 8; - case Size::Size_16_16_16: + case Size::Size_R16_G16_B16: return 6; - case Size::Size_8_8_8_8: + case Size::Size_R8_G8_B8_A8: + case Size::Size_X8_B8_G8_R8: return 4; - case Size::Size_16_16: + case Size::Size_R16_G16: return 4; - case Size::Size_32: + case Size::Size_R32: return 4; - case Size::Size_8_8_8: + case Size::Size_R8_G8_B8: return 3; - case Size::Size_8_8: + case Size::Size_R8_G8: + case Size::Size_G8_R8: return 2; - case Size::Size_16: + case Size::Size_R16: return 2; - case Size::Size_8: + case Size::Size_R8: + case Size::Size_A8: return 1; - case Size::Size_10_10_10_2: + case Size::Size_A2_B10_G10_R10: return 4; - case Size::Size_11_11_10: + case Size::Size_B10_G11_R11: return 4; default: ASSERT(false); @@ -245,34 +678,36 @@ public: std::string SizeString() const { switch (size) { - case Size::Size_32_32_32_32: + case Size::Size_R32_G32_B32_A32: return "32_32_32_32"; - case Size::Size_32_32_32: + case Size::Size_R32_G32_B32: return "32_32_32"; - case Size::Size_16_16_16_16: + case Size::Size_R16_G16_B16_A16: return "16_16_16_16"; - case Size::Size_32_32: + case Size::Size_R32_G32: return "32_32"; - case Size::Size_16_16_16: + case Size::Size_R16_G16_B16: return "16_16_16"; - case Size::Size_8_8_8_8: + case Size::Size_R8_G8_B8_A8: return "8_8_8_8"; - case Size::Size_16_16: + case Size::Size_R16_G16: return "16_16"; - case Size::Size_32: + case Size::Size_R32: return "32"; - case Size::Size_8_8_8: + case Size::Size_R8_G8_B8: return "8_8_8"; - case Size::Size_8_8: + case Size::Size_R8_G8: + case Size::Size_G8_R8: return "8_8"; - case Size::Size_16: + case Size::Size_R16: return "16"; - case Size::Size_8: + case Size::Size_R8: + case Size::Size_A8: return "8"; - case Size::Size_10_10_10_2: - return "10_10_10_2"; - case Size::Size_11_11_10: - return "11_11_10"; + case Size::Size_A2_B10_G10_R10: + return "2_10_10_10"; + case Size::Size_B10_G11_R11: + return "10_11_12"; default: ASSERT(false); return {}; @@ -281,17 +716,19 @@ public: std::string TypeString() const { switch (type) { - case Type::SignedNorm: + case Type::UnusedEnumDoNotUseBecauseItWillGoAway: + return "Unused"; + case Type::SNorm: return "SNORM"; - case Type::UnsignedNorm: + case Type::UNorm: return "UNORM"; - case Type::SignedInt: + case Type::SInt: return "SINT"; - case Type::UnsignedInt: + case Type::UInt: return "UINT"; - case Type::UnsignedScaled: + case Type::UScaled: return "USCALED"; - case Type::SignedScaled: + case Type::SScaled: return "SSCALED"; case Type::Float: return "FLOAT"; @@ -301,7 +738,7 @@ public: } bool IsNormalized() const { - return (type == Type::SignedNorm) || (type == Type::UnsignedNorm); + return (type == Type::SNorm) || (type == Type::UNorm); } bool IsValid() const { @@ -312,6 +749,7 @@ public: return hex < other.hex; } }; + static_assert(sizeof(VertexAttribute) == 0x4); struct MsaaSampleLocation { union { @@ -342,9 +780,96 @@ public: } }; - enum class DepthMode : u32 { - MinusOneToOne = 0, - ZeroToOne = 1, + struct MultisampleCoverageToColor { + union { + BitField<0, 1, u32> enable; + BitField<4, 3, u32> target; + }; + }; + + struct DecompressZetaSurface { + union { + BitField<0, 1, u32> z_enable; + BitField<4, 1, u32> stencil_enable; + }; + }; + + struct ZetaSparse { + enum class UnmappedCompare : u32 { + Unmapped = 0, + FailAlways = 1, + }; + union { + BitField<0, 1, u32> enable; + BitField<1, 1, UnmappedCompare> unmapped_compare; + }; + }; + + struct RtControl { + union { + BitField<0, 4, u32> count; + BitField<4, 3, u32> target0; + BitField<7, 3, u32> target1; + BitField<10, 3, u32> target2; + BitField<13, 3, u32> target3; + BitField<16, 3, u32> target4; + BitField<19, 3, u32> target5; + BitField<22, 3, u32> target6; + BitField<25, 3, u32> target7; + }; + + u32 Map(std::size_t index) const { + const std::array<u32, NumRenderTargets> maps{target0, target1, target2, target3, + target4, target5, target6, target7}; + ASSERT(index < maps.size()); + return maps[index]; + } + }; + + struct CompressionThresholdSamples { + u32 samples; + + u32 Samples() { + if (samples == 0) { + return 0; + } + return 1 << (samples - 1); + } + }; + + struct PixelShaderInterlockControl { + enum class TileMode : u32 { + NoConflictDetect = 0, + DetectSampleConflict = 1, + DetectPixelConflict = 2, + }; + enum class TileSize : u32 { + Size_16x16 = 0, + Size_8x8 = 1, + }; + enum class FragmentOrder : u32 { + FragmentOrdered = 0, + FragmentUnordered = 1, + }; + union { + BitField<0, 2, TileMode> tile_mode; + BitField<2, 1, TileSize> tile_size; + BitField<3, 1, FragmentOrder> fragment_order; + }; + }; + + struct ZetaSize { + enum class DimensionControl : u32 { + DepthDefinesArray = 0, + ArraySizeOne = 1, + }; + + u32 width; + u32 height; + union { + BitField<0, 16, u32> depth; + BitField<16, 1, DimensionControl> dim_control; + }; }; enum class PrimitiveTopology : u32 { @@ -358,15 +883,21 @@ public: Quads = 0x7, QuadStrip = 0x8, Polygon = 0x9, - LinesAdjacency = 0xa, - LineStripAdjacency = 0xb, - TrianglesAdjacency = 0xc, - TriangleStripAdjacency = 0xd, - Patches = 0xe, + LinesAdjacency = 0xA, + LineStripAdjacency = 0xB, + TrianglesAdjacency = 0xC, + TriangleStripAdjacency = 0xD, + Patches = 0xE, + }; + + struct VertexArray { + union { + BitField<0, 16, u32> start; + BitField<16, 12, u32> count; + BitField<28, 3, PrimitiveTopology> topology; + }; }; - // Constants as from NVC0_3D_UNK1970_D3D - // https://gitlab.freedesktop.org/mesa/mesa/-/blob/main/src/gallium/drivers/nouveau/nvc0/nvc0_3d.xml.h#L1598 enum class PrimitiveTopologyOverride : u32 { None = 0x0, Points = 0x1, @@ -374,11 +905,32 @@ public: LineStrip = 0x3, Triangles = 0x4, TriangleStrip = 0x5, - LinesAdjacency = 0xa, - LineStripAdjacency = 0xb, - TrianglesAdjacency = 0xc, - TriangleStripAdjacency = 0xd, - Patches = 0xe, + LinesAdjacency = 0xA, + LineStripAdjacency = 0xB, + TrianglesAdjacency = 0xC, + TriangleStripAdjacency = 0xD, + Patches = 0xE, + + LegacyPoints = 0x1001, + LegacyIndexedLines = 0x1002, + LegacyIndexedTriangles = 0x1003, + LegacyLines = 0x100F, + LegacyLineStrip = 0x1010, + LegacyIndexedLineStrip = 0x1011, + LegacyTriangles = 0x1012, + LegacyTriangleStrip = 0x1013, + LegacyIndexedTriangleStrip = 0x1014, + LegacyTriangleFan = 0x1015, + LegacyIndexedTriangleFan = 0x1016, + LegacyTriangleFanImm = 0x1017, + LegacyLinesImm = 0x1018, + LegacyIndexedTriangles2 = 0x101A, + LegacyIndexedLines2 = 0x101B, + }; + + enum class DepthMode : u32 { + MinusOneToOne = 0, + ZeroToOne = 1, }; enum class IndexFormat : u32 { @@ -388,183 +940,143 @@ public: }; enum class ComparisonOp : u32 { - // These values are used by Nouveau and most games, they correspond to the OpenGL token - // values for these operations. - Never = 0x200, - Less = 0x201, - Equal = 0x202, - LessEqual = 0x203, - Greater = 0x204, - NotEqual = 0x205, - GreaterEqual = 0x206, - Always = 0x207, - - // These values are used by some games, they seem to be NV04 values. - NeverOld = 1, - LessOld = 2, - EqualOld = 3, - LessEqualOld = 4, - GreaterOld = 5, - NotEqualOld = 6, - GreaterEqualOld = 7, - AlwaysOld = 8, - }; - - enum class LogicOperation : u32 { - Clear = 0x1500, - And = 0x1501, - AndReverse = 0x1502, - Copy = 0x1503, - AndInverted = 0x1504, - NoOp = 0x1505, - Xor = 0x1506, - Or = 0x1507, - Nor = 0x1508, - Equiv = 0x1509, - Invert = 0x150A, - OrReverse = 0x150B, - CopyInverted = 0x150C, - OrInverted = 0x150D, - Nand = 0x150E, - Set = 0x150F, - }; - - enum class StencilOp : u32 { - Keep = 1, - Zero = 2, - Replace = 3, - Incr = 4, - Decr = 5, - Invert = 6, - IncrWrap = 7, - DecrWrap = 8, - KeepOGL = 0x1E00, - ZeroOGL = 0, - ReplaceOGL = 0x1E01, - IncrOGL = 0x1E02, - DecrOGL = 0x1E03, - InvertOGL = 0x150A, - IncrWrapOGL = 0x8507, - DecrWrapOGL = 0x8508, - }; - - enum class CounterReset : u32 { - SampleCnt = 0x01, - Unk02 = 0x02, - Unk03 = 0x03, - Unk04 = 0x04, - EmittedPrimitives = 0x10, // Not tested - Unk11 = 0x11, - Unk12 = 0x12, - Unk13 = 0x13, - Unk15 = 0x15, - Unk16 = 0x16, - Unk17 = 0x17, - Unk18 = 0x18, - Unk1A = 0x1A, - Unk1B = 0x1B, - Unk1C = 0x1C, - Unk1D = 0x1D, - Unk1E = 0x1E, - GeneratedPrimitives = 0x1F, + Never_D3D = 1, + Less_D3D = 2, + Equal_D3D = 3, + LessEqual_D3D = 4, + Greater_D3D = 5, + NotEqual_D3D = 6, + GreaterEqual_D3D = 7, + Always_D3D = 8, + + Never_GL = 0x200, + Less_GL = 0x201, + Equal_GL = 0x202, + LessEqual_GL = 0x203, + Greater_GL = 0x204, + NotEqual_GL = 0x205, + GreaterEqual_GL = 0x206, + Always_GL = 0x207, + }; + + enum class ClearReport : u32 { + ZPassPixelCount = 0x01, + ZCullStats = 0x02, + StreamingPrimitvesNeededMinusSucceeded = 0x03, + AlphaBetaClocks = 0x04, + StreamingPrimitivesSucceeded = 0x10, + StreamingPrimitivesNeeded = 0x11, + VerticesGenerated = 0x12, + PrimitivesGenerated = 0x13, + VertexShaderInvocations = 0x15, + TessellationInitInvocations = 0x16, + TessellationShaderInvocations = 0x17, + TessellationShaderPrimitivesGenerated = 0x18, + GeometryShaderInvocations = 0x1A, + GeometryShaderPrimitivesGenerated = 0x1B, + ClipperInvocations = 0x1C, + ClipperPrimitivesGenerated = 0x1D, + PixelShaderInvocations = 0x1E, + VtgPrimitivesOut = 0x1F, }; enum class FrontFace : u32 { - ClockWise = 0x0900, - CounterClockWise = 0x0901, + ClockWise = 0x900, + CounterClockWise = 0x901, }; enum class CullFace : u32 { - Front = 0x0404, - Back = 0x0405, - FrontAndBack = 0x0408, + Front = 0x404, + Back = 0x405, + FrontAndBack = 0x408, }; struct Blend { enum class Equation : u32 { - Add = 1, - Subtract = 2, - ReverseSubtract = 3, - Min = 4, - Max = 5, - - // These values are used by Nouveau and some games. - AddGL = 0x8006, - MinGL = 0x8007, - MaxGL = 0x8008, - SubtractGL = 0x800a, - ReverseSubtractGL = 0x800b + Add_D3D = 1, + Subtract_D3D = 2, + ReverseSubtract_D3D = 3, + Min_D3D = 4, + Max_D3D = 5, + + Add_GL = 0x8006, + Min_GL = 0x8007, + Max_GL = 0x8008, + Subtract_GL = 0x800A, + ReverseSubtract_GL = 0x800B }; enum class Factor : u32 { - Zero = 0x1, - One = 0x2, - SourceColor = 0x3, - OneMinusSourceColor = 0x4, - SourceAlpha = 0x5, - OneMinusSourceAlpha = 0x6, - DestAlpha = 0x7, - OneMinusDestAlpha = 0x8, - DestColor = 0x9, - OneMinusDestColor = 0xa, - SourceAlphaSaturate = 0xb, - Source1Color = 0x10, - OneMinusSource1Color = 0x11, - Source1Alpha = 0x12, - OneMinusSource1Alpha = 0x13, - ConstantColor = 0x61, - OneMinusConstantColor = 0x62, - ConstantAlpha = 0x63, - OneMinusConstantAlpha = 0x64, - - // These values are used by Nouveau and some games. - ZeroGL = 0x4000, - OneGL = 0x4001, - SourceColorGL = 0x4300, - OneMinusSourceColorGL = 0x4301, - SourceAlphaGL = 0x4302, - OneMinusSourceAlphaGL = 0x4303, - DestAlphaGL = 0x4304, - OneMinusDestAlphaGL = 0x4305, - DestColorGL = 0x4306, - OneMinusDestColorGL = 0x4307, - SourceAlphaSaturateGL = 0x4308, - ConstantColorGL = 0xc001, - OneMinusConstantColorGL = 0xc002, - ConstantAlphaGL = 0xc003, - OneMinusConstantAlphaGL = 0xc004, - Source1ColorGL = 0xc900, - OneMinusSource1ColorGL = 0xc901, - Source1AlphaGL = 0xc902, - OneMinusSource1AlphaGL = 0xc903, + Zero_D3D = 0x1, + One_D3D = 0x2, + SourceColor_D3D = 0x3, + OneMinusSourceColor_D3D = 0x4, + SourceAlpha_D3D = 0x5, + OneMinusSourceAlpha_D3D = 0x6, + DestAlpha_D3D = 0x7, + OneMinusDestAlpha_D3D = 0x8, + DestColor_D3D = 0x9, + OneMinusDestColor_D3D = 0xA, + SourceAlphaSaturate_D3D = 0xB, + BothSourceAlpha_D3D = 0xC, + OneMinusBothSourceAlpha_D3D = 0xD, + BlendFactor_D3D = 0xE, + OneMinusBlendFactor_D3D = 0xF, + Source1Color_D3D = 0x10, + OneMinusSource1Color_D3D = 0x11, + Source1Alpha_D3D = 0x12, + OneMinusSource1Alpha_D3D = 0x13, + + Zero_GL = 0x4000, + One_GL = 0x4001, + SourceColor_GL = 0x4300, + OneMinusSourceColor_GL = 0x4301, + SourceAlpha_GL = 0x4302, + OneMinusSourceAlpha_GL = 0x4303, + DestAlpha_GL = 0x4304, + OneMinusDestAlpha_GL = 0x4305, + DestColor_GL = 0x4306, + OneMinusDestColor_GL = 0x4307, + SourceAlphaSaturate_GL = 0x4308, + ConstantColor_GL = 0xC001, + OneMinusConstantColor_GL = 0xC002, + ConstantAlpha_GL = 0xC003, + OneMinusConstantAlpha_GL = 0xC004, + Source1Color_GL = 0xC900, + OneMinusSource1Color_GL = 0xC901, + Source1Alpha_GL = 0xC902, + OneMinusSource1Alpha_GL = 0xC903, }; u32 separate_alpha; - Equation equation_rgb; - Factor factor_source_rgb; - Factor factor_dest_rgb; - Equation equation_a; - Factor factor_source_a; - Factor factor_dest_a; - INSERT_PADDING_WORDS_NOINIT(1); - }; - - enum class TessellationPrimitive : u32 { - Isolines = 0, - Triangles = 1, - Quads = 2, + Equation color_op; + Factor color_source; + Factor color_dest; + Equation alpha_op; + Factor alpha_source; + u32 enable_global_color_key; + Factor alpha_dest; + + u32 single_rop_control_enable; + u32 enable[NumRenderTargets]; }; - enum class TessellationSpacing : u32 { - Equal = 0, - FractionalOdd = 1, - FractionalEven = 2, + struct BlendPerTarget { + u32 separate_alpha; + Blend::Equation color_op; + Blend::Factor color_source; + Blend::Factor color_dest; + Blend::Equation alpha_op; + Blend::Factor alpha_source; + Blend::Factor alpha_dest; + INSERT_PADDING_BYTES_NOINIT(0x4); }; + static_assert(sizeof(BlendPerTarget) == 0x20); enum class PolygonMode : u32 { - Point = 0x1b00, - Line = 0x1b01, - Fill = 0x1b02, + Point = 0x1B00, + Line = 0x1B01, + Fill = 0x1B02, }; enum class ShadowRamControl : u32 { @@ -589,18 +1101,22 @@ public: NegativeW = 7, }; - enum class SamplerIndex : u32 { + enum class SamplerBinding : u32 { Independently = 0, - ViaHeaderIndex = 1, + ViaHeaderBinding = 1, }; struct TileMode { + enum class DimensionControl : u32 { + DepthDefinesArray = 0, + DepthDefinesDepth = 1, + }; union { BitField<0, 4, u32> block_width; BitField<4, 4, u32> block_height; BitField<8, 4, u32> block_depth; BitField<12, 1, u32> is_pitch_linear; - BitField<16, 1, u32> is_3d; + BitField<16, 1, DimensionControl> dim_control; }; }; static_assert(sizeof(TileMode) == 4); @@ -616,23 +1132,25 @@ public: BitField<0, 16, u32> depth; BitField<16, 1, u32> volume; }; - u32 layer_stride; + u32 array_pitch; u32 base_layer; - INSERT_PADDING_WORDS_NOINIT(7); + u32 mark_ieee_clean; + INSERT_PADDING_BYTES_NOINIT(0x18); GPUVAddr Address() const { return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | address_low); } }; + static_assert(sizeof(RenderTargetConfig) == 0x40); struct ColorMask { union { u32 raw; - BitField<0, 4, u32> R; - BitField<4, 4, u32> G; - BitField<8, 4, u32> B; - BitField<12, 4, u32> A; + BitField<0, 1, u32> R; + BitField<4, 1, u32> G; + BitField<8, 1, u32> B; + BitField<12, 1, u32> A; }; }; @@ -643,6 +1161,7 @@ public: f32 translate_x; f32 translate_y; f32 translate_z; + union { u32 raw; BitField<0, 3, ViewportSwizzle> x; @@ -650,7 +1169,11 @@ public: BitField<8, 3, ViewportSwizzle> z; BitField<12, 3, ViewportSwizzle> w; } swizzle; - INSERT_PADDING_WORDS_NOINIT(1); + + union { + BitField<0, 5, u32> x; + BitField<8, 5, u32> y; + } snap_grid_precision; Common::Rectangle<f32> GetRect() const { return { @@ -677,21 +1200,14 @@ public: return translate_y + std::fabs(scale_y) - GetY(); } }; + static_assert(sizeof(ViewportTransform) == 0x20); - struct ScissorTest { - u32 enable; - union { - BitField<0, 16, u32> min_x; - BitField<16, 16, u32> max_x; + struct Viewport { + enum class PixelCenter : u32 { + HalfIntegers = 0, + Integers = 1, }; - union { - BitField<0, 16, u32> min_y; - BitField<16, 16, u32> max_y; - }; - u32 fill; - }; - struct ViewPort { union { BitField<0, 16, u32> x; BitField<16, 16, u32> width; @@ -703,726 +1219,1820 @@ public: float depth_range_near; float depth_range_far; }; + static_assert(sizeof(Viewport) == 0x10); - struct TransformFeedbackBinding { - u32 buffer_enable; - u32 address_high; - u32 address_low; - s32 buffer_size; - s32 buffer_offset; - INSERT_PADDING_WORDS_NOINIT(3); - - GPUVAddr Address() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | - address_low); - } + struct Window { + union { + BitField<0, 16, u32> x_min; + BitField<16, 16, u32> x_max; + }; + union { + BitField<0, 16, u32> y_min; + BitField<16, 16, u32> y_max; + }; }; - static_assert(sizeof(TransformFeedbackBinding) == 32); + static_assert(sizeof(Window) == 0x8); - struct TransformFeedbackLayout { - u32 stream; - u32 varying_count; - u32 stride; - INSERT_PADDING_WORDS_NOINIT(1); + struct ClipIdExtent { + union { + BitField<0, 16, u32> x; + BitField<16, 16, u32> width; + }; + union { + BitField<0, 16, u32> y; + BitField<16, 16, u32> height; + }; + }; + static_assert(sizeof(ClipIdExtent) == 0x8); + + enum class VisibleCallLimit : u32 { + Limit0 = 0, + Limit1 = 1, + Limit2 = 2, + Limit4 = 3, + Limit8 = 4, + Limit16 = 5, + Limit32 = 6, + Limit64 = 7, + Limit128 = 8, + None = 15, }; - static_assert(sizeof(TransformFeedbackLayout) == 16); - - bool IsShaderConfigEnabled(std::size_t index) const { - // The VertexB is always enabled. - if (index == static_cast<std::size_t>(Regs::ShaderProgram::VertexB)) { - return true; - } - return shader_config[index].enable != 0; - } - - bool IsShaderConfigEnabled(Regs::ShaderProgram type) const { - return IsShaderConfigEnabled(static_cast<std::size_t>(type)); - } - - union { - struct { - INSERT_PADDING_WORDS_NOINIT(0x44); - u32 wait_for_idle; + struct StatisticsCounter { + union { + BitField<0, 1, u32> da_vertices; + BitField<1, 1, u32> da_primitives; + BitField<2, 1, u32> vs_invocations; + BitField<3, 1, u32> gs_invocations; + BitField<4, 1, u32> gs_primitives; + BitField<5, 1, u32> streaming_primitives_succeeded; + BitField<6, 1, u32> streaming_primitives_needed; + BitField<7, 1, u32> clipper_invocations; + BitField<8, 1, u32> clipper_primitives; + BitField<9, 1, u32> ps_invocations; + BitField<11, 1, u32> ti_invocations; + BitField<12, 1, u32> ts_invocations; + BitField<13, 1, u32> ts_primitives; + BitField<14, 1, u32> total_streaming_primitives_needed_succeeded; + BitField<10, 1, u32> vtg_primitives_out; + BitField<15, 1, u32> alpha_beta_clocks; + }; + }; - struct { - u32 upload_address; - u32 data; - u32 entry; - u32 bind; - } macros; + struct ClearRect { + union { + BitField<0, 16, u32> x_min; + BitField<16, 16, u32> x_max; + }; + union { + BitField<0, 16, u32> y_min; + BitField<16, 16, u32> y_max; + }; + }; - ShadowRamControl shadow_ram_control; + struct VertexBuffer { + u32 first; + u32 count; + }; - INSERT_PADDING_WORDS_NOINIT(0x16); + struct InvalidateShaderCacheNoWFI { + union { + BitField<0, 1, u32> instruction; + BitField<4, 1, u32> global_data; + BitField<12, 1, u32> constant; + }; + }; - Upload::Registers upload; - struct { - union { - BitField<0, 1, u32> linear; - }; - } exec_upload; + struct ZCullSerialization { + enum class Applied : u32 { + Always = 0, + LateZ = 1, + OutOfGamutZ = 2, + LateZOrOutOfGamutZ = 3, + }; + union { + BitField<0, 1, u32> enable; + BitField<4, 2, Applied> applied; + }; + }; - u32 data_upload; + struct ZCullDirFormat { + enum class Zdir : u32 { + Less = 0, + Greater = 1, + }; + enum class Zformat : u32 { + MSB = 0, + FP = 1, + Ztrick = 2, + Zf32 = 3, + }; - INSERT_PADDING_WORDS_NOINIT(0x16); + union { + BitField<0, 16, Zdir> dir; + BitField<16, 16, Zformat> format; + }; + }; - u32 force_early_fragment_tests; + struct IteratedBlend { + union { + BitField<0, 1, u32> enable; + BitField<1, 1, u32> enable_alpha; + }; + u32 pass_count; + }; - INSERT_PADDING_WORDS_NOINIT(0x2D); + struct ZCullCriterion { + enum class Sfunc : u32 { + Never = 0, + Less = 1, + Equal = 2, + LessOrEqual = 3, + Greater = 4, + NotEqual = 5, + GreaterOrEqual = 6, + Always = 7, + }; - struct { - union { - BitField<0, 16, u32> sync_point; - BitField<16, 1, u32> unknown; - BitField<20, 1, u32> increment; - }; - } sync_info; + union { + BitField<0, 8, Sfunc> sfunc; + BitField<8, 1, u32> no_invalidate; + BitField<9, 1, u32> force_match; + BitField<16, 8, u32> sref; + BitField<24, 8, u32> smask; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x15); + struct LoadIteratedBlend { + enum class Test : u32 { + False = 0, + True = 1, + Equal = 2, + NotEqual = 3, + LessThan = 4, + LessOrEqual = 5, + Greater = 6, + GreaterOrEqual = 7, + }; + enum class Operation : u32 { + AddProducts = 0, + SubProducts = 1, + Min = 2, + Max = 3, + Reciprocal = 4, + Add = 5, + Sub = 6, + }; + enum class OperandA : u32 { + SrcRGB = 0, + DstRGB = 1, + SrcAAA = 2, + DstAAA = 3, + Temp0_RGB = 4, + Temp1_RGB = 5, + Temp2_RGB = 6, + PBR_RGB = 7, + }; + enum class OperandB : u32 { + Zero = 0, + One = 1, + SrcRGB = 2, + SrcAAA = 3, + OneMinusSrcAAA = 4, + DstRGB = 5, + DstAAA = 6, + OneMinusDstAAA = 7, + Temp0_RGB = 9, + Temp1_RGB = 10, + Temp2_RGB = 11, + PBR_RGB = 12, + ConstRGB = 13, + ZeroATimesB = 14, + }; + enum class Swizzle : u32 { + RGB = 0, + GBR = 1, + RRR = 2, + GGG = 3, + BBB = 4, + RToA = 5, + }; + enum class WriteMask : u32 { + RGB = 0, + ROnly = 1, + GOnly = 2, + BOnly = 3, + }; + enum class Pass : u32 { + Temp0 = 0, + Temp1 = 1, + Temp2 = 2, + None = 3, + }; - union { - BitField<0, 2, TessellationPrimitive> prim; - BitField<4, 2, TessellationSpacing> spacing; - BitField<8, 1, u32> cw; - BitField<9, 1, u32> connected; - } tess_mode; + u32 instruction_ptr; + union { + BitField<0, 3, Test> test; + BitField<3, 3, Operation> operation; + BitField<6, 3, u32> const_input; + BitField<9, 3, OperandA> operand_a; + BitField<12, 4, OperandB> operand_b; + BitField<16, 3, OperandA> operand_c; + BitField<19, 4, OperandB> operand_d; + BitField<23, 3, Swizzle> output_swizzle; + BitField<26, 2, WriteMask> output_mask; + BitField<28, 2, Pass> output_pass; + BitField<31, 1, u32> test_enabled; + }; + }; - std::array<f32, 4> tess_level_outer; - std::array<f32, 2> tess_level_inner; + struct ScissorTest { + u32 enable; + union { + BitField<0, 16, u32> min_x; + BitField<16, 16, u32> max_x; + }; + union { + BitField<0, 16, u32> min_y; + BitField<16, 16, u32> max_y; + }; + INSERT_PADDING_BYTES_NOINIT(0x4); + }; + static_assert(sizeof(ScissorTest) == 0x10); - INSERT_PADDING_WORDS_NOINIT(0x10); + struct VPCPerf { + union { + BitField<0, 8, u32> culled_small_lines; + BitField<8, 8, u32> culled_small_triangles; + BitField<16, 8, u32> nonculled_lines_and_points; + BitField<24, 8, u32> nonculled_triangles; + }; + }; - u32 rasterize_enable; + struct ConstantColorRendering { + u32 enabled; + u32 red; + u32 green; + u32 blue; + u32 alpha; + }; - std::array<TransformFeedbackBinding, NumTransformFeedbackBuffers> tfb_bindings; + struct VertexStreamSubstitute { + u32 address_high; + u32 address_low; - INSERT_PADDING_WORDS_NOINIT(0xC0); + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - std::array<TransformFeedbackLayout, NumTransformFeedbackBuffers> tfb_layouts; + struct VTGWarpWatermarks { + union { + BitField<0, 16, u32> low; + BitField<16, 16, u32> high; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x1); + struct SampleMask { + struct Target { + union { + BitField<0, 1, u32> raster_out; + BitField<4, 1, u32> color_target; + }; + u32 target; + }; + struct Pos { + u32 x0_y0; + u32 x1_y0; + u32 x0_y1; + u32 x1_y1; + }; + }; - u32 tfb_enabled; + enum class NonMultisampledZ : u32 { + PerSample = 0, + PixelCenter = 1, + }; - INSERT_PADDING_WORDS_NOINIT(0x2E); + enum class TIRMode : u32 { + Disabled = 0, + RasterNTargetM = 1, + }; - std::array<RenderTargetConfig, NumRenderTargets> rt; + enum class AntiAliasRaster : u32 { + Mode1x1 = 0, + Mode2x2 = 2, + Mode4x2_D3D = 4, + Mode2x1_D3D = 5, + Mode4x4 = 6, + }; - std::array<ViewportTransform, NumViewports> viewport_transform; + struct SurfaceClipIDMemory { + u32 address_high; + u32 address_low; - std::array<ViewPort, NumViewports> viewports; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - INSERT_PADDING_WORDS_NOINIT(0x1D); + struct TIRModulation { + enum class Component : u32 { + None = 0, + RGB = 1, + AlphaOnly = 2, + RGBA = 3, + }; + enum class Function : u32 { + Linear = 0, + Table = 1, + }; + Component component; + Function function; + }; - struct { - u32 first; - u32 count; - } vertex_buffer; + struct Zeta { + u32 address_high; + u32 address_low; + Tegra::DepthFormat format; + TileMode tile_mode; + u32 array_pitch; - DepthMode depth_mode; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - float clear_color[4]; - float clear_depth; + struct SurfaceClip { + union { + BitField<0, 16, u32> x; + BitField<16, 16, u32> width; + }; + union { + BitField<0, 16, u32> y; + BitField<16, 16, u32> height; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x3); + enum class L2CacheControlPolicy : u32 { + First = 0, + Normal = 1, + Last = 2, + }; - s32 clear_stencil; + struct L2CacheVAFRequests { + union { + BitField<0, 1, u32> system_memory_volatile; + BitField<4, 2, L2CacheControlPolicy> policy; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x2); + enum class ViewportMulticast : u32 { + ViewportOrder = 0, + PrimitiveOrder = 1, + }; - PolygonMode polygon_mode_front; - PolygonMode polygon_mode_back; + struct TIRModulationCoeff { + union { + BitField<0, 8, u32> table_v0; + BitField<8, 8, u32> table_v1; + BitField<16, 8, u32> table_v2; + BitField<24, 8, u32> table_v3; + }; + }; + static_assert(sizeof(TIRModulationCoeff) == 0x4); - INSERT_PADDING_WORDS_NOINIT(0x3); + struct ReduceColorThreshold { + union { + BitField<0, 8, u32> all_hit_once; + BitField<16, 8, u32> all_covered; + }; + }; - u32 polygon_offset_point_enable; - u32 polygon_offset_line_enable; - u32 polygon_offset_fill_enable; + struct ClearControl { + union { + BitField<0, 1, u32> respect_stencil_mask; + BitField<4, 1, u32> use_clear_rect; + BitField<8, 1, u32> use_scissor; + BitField<12, 1, u32> use_viewport_clip0; + }; + }; - u32 patch_vertices; + struct L2CacheRopNonInterlockedReads { + union { + BitField<4, 2, L2CacheControlPolicy> policy; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x4); + struct VertexOutputAttributeSkipMasks { + struct Attributes { + union { + BitField<0, 1, u32> attribute0_comp0; + BitField<1, 1, u32> attribute0_comp1; + BitField<2, 1, u32> attribute0_comp2; + BitField<3, 1, u32> attribute0_comp3; + BitField<4, 1, u32> attribute1_comp0; + BitField<5, 1, u32> attribute1_comp1; + BitField<6, 1, u32> attribute1_comp2; + BitField<7, 1, u32> attribute1_comp3; + BitField<8, 1, u32> attribute2_comp0; + BitField<9, 1, u32> attribute2_comp1; + BitField<10, 1, u32> attribute2_comp2; + BitField<11, 1, u32> attribute2_comp3; + BitField<12, 1, u32> attribute3_comp0; + BitField<13, 1, u32> attribute3_comp1; + BitField<14, 1, u32> attribute3_comp2; + BitField<15, 1, u32> attribute3_comp3; + BitField<16, 1, u32> attribute4_comp0; + BitField<17, 1, u32> attribute4_comp1; + BitField<18, 1, u32> attribute4_comp2; + BitField<19, 1, u32> attribute4_comp3; + BitField<20, 1, u32> attribute5_comp0; + BitField<21, 1, u32> attribute5_comp1; + BitField<22, 1, u32> attribute5_comp2; + BitField<23, 1, u32> attribute5_comp3; + BitField<24, 1, u32> attribute6_comp0; + BitField<25, 1, u32> attribute6_comp1; + BitField<26, 1, u32> attribute6_comp2; + BitField<27, 1, u32> attribute6_comp3; + BitField<28, 1, u32> attribute7_comp0; + BitField<29, 1, u32> attribute7_comp1; + BitField<30, 1, u32> attribute7_comp2; + BitField<31, 1, u32> attribute7_comp3; + }; + }; - u32 fragment_barrier; + std::array<Attributes, 2> a; + std::array<Attributes, 2> b; + }; - INSERT_PADDING_WORDS_NOINIT(0x7); + struct TIRControl { + union { + BitField<0, 1, u32> z_pass_pixel_count_use_raster_samples; + BitField<4, 1, u32> alpha_coverage_use_raster_samples; + BitField<1, 1, u32> reduce_coverage; + }; + }; - std::array<ScissorTest, NumViewports> scissor_test; + enum class FillViaTriangleMode : u32 { + Disabled = 0, + FillAll = 1, + FillBoundingBox = 2, + }; - INSERT_PADDING_WORDS_NOINIT(0x15); + struct PsTicketDispenserValue { + union { + BitField<0, 8, u32> index; + BitField<8, 16, u32> value; + }; + }; - s32 stencil_back_func_ref; - u32 stencil_back_mask; - u32 stencil_back_func_mask; + struct RegisterWatermarks { + union { + BitField<0, 16, u32> low; + BitField<16, 16, u32> high; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x5); + enum class InvalidateCacheLines : u32 { + All = 0, + One = 1, + }; - u32 invalidate_texture_data_cache; + struct InvalidateTextureDataCacheNoWfi { + union { + BitField<0, 1, InvalidateCacheLines> lines; + BitField<4, 22, u32> tag; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x1); + struct ZCullRegionEnable { + union { + BitField<0, 1, u32> enable_z; + BitField<4, 1, u32> enable_stencil; + BitField<1, 1, u32> rect_clear; + BitField<2, 1, u32> use_rt_array_index; + BitField<5, 16, u32> rt_array_index; + BitField<3, 1, u32> make_conservative; + }; + }; - u32 tiled_cache_barrier; + enum class FillMode : u32 { + Point = 1, + Wireframe = 2, + Solid = 3, + }; - INSERT_PADDING_WORDS_NOINIT(0x4); + enum class ShadeMode : u32 { + Flat = 0x1, + Gouraud = 0x2, + GL_Flat = 0x1D00, + GL_Smooth = 0x1D01, + }; - u32 color_mask_common; + enum class AlphaToCoverageDither : u32 { + Footprint_1x1 = 0, + Footprint_2x2 = 1, + Footprint_1x1_Virtual = 2, + }; - INSERT_PADDING_WORDS_NOINIT(0x2); + struct InlineIndex4x8Align { + union { + BitField<0, 30, u32> count; + BitField<30, 2, u32> start; + }; + }; - f32 depth_bounds[2]; + struct InlineIndex4x8Index { + union { + BitField<0, 8, u32> index0; + BitField<8, 8, u32> index1; + BitField<16, 8, u32> index2; + BitField<24, 8, u32> index3; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x2); + enum class D3DCullMode : u32 { + None = 0, + CW = 1, + CCW = 2, + }; - u32 rt_separate_frag_data; + struct BlendColor { + f32 r; + f32 g; + f32 b; + f32 a; + }; - INSERT_PADDING_WORDS_NOINIT(0x1); + struct StencilOp { + enum class Op : u32 { + Keep_D3D = 1, + Zero_D3D = 2, + Replace_D3D = 3, + IncrSaturate_D3D = 4, + DecrSaturate_D3D = 5, + Invert_D3D = 6, + Incr_D3D = 7, + Decr_D3D = 8, + + Keep_GL = 0x1E00, + Zero_GL = 0, + Replace_GL = 0x1E01, + IncrSaturate_GL = 0x1E02, + DecrSaturate_GL = 0x1E03, + Invert_GL = 0x150A, + Incr_GL = 0x8507, + Decr_GL = 0x8508, + }; - u32 multisample_raster_enable; - u32 multisample_raster_samples; - std::array<u32, 4> multisample_sample_mask; + Op fail; + Op zfail; + Op zpass; + ComparisonOp func; + }; - INSERT_PADDING_WORDS_NOINIT(0x5); + struct PsSaturate { + // Opposite of DepthMode + enum class Depth : u32 { + ZeroToOne = 0, + MinusOneToOne = 1, + }; - struct { - u32 address_high; - u32 address_low; - Tegra::DepthFormat format; - TileMode tile_mode; - u32 layer_stride; + union { + BitField<0, 1, u32> output0_enable; + BitField<1, 1, Depth> output0_range; + BitField<4, 1, u32> output1_enable; + BitField<5, 1, Depth> output1_range; + BitField<8, 1, u32> output2_enable; + BitField<9, 1, Depth> output2_range; + BitField<12, 1, u32> output3_enable; + BitField<13, 1, Depth> output3_range; + BitField<16, 1, u32> output4_enable; + BitField<17, 1, Depth> output4_range; + BitField<20, 1, u32> output5_enable; + BitField<21, 1, Depth> output5_range; + BitField<24, 1, u32> output6_enable; + BitField<25, 1, Depth> output6_range; + BitField<28, 1, u32> output7_enable; + BitField<29, 1, Depth> output7_range; + }; - GPUVAddr Address() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | - address_low); - } - } zeta; + bool AnyEnabled() const { + return output0_enable || output1_enable || output2_enable || output3_enable || + output4_enable || output5_enable || output6_enable || output7_enable; + } + }; - struct { - union { - BitField<0, 16, u32> x; - BitField<16, 16, u32> width; - }; - union { - BitField<0, 16, u32> y; - BitField<16, 16, u32> height; - }; - } render_area; + struct WindowOrigin { + enum class Mode : u32 { + UpperLeft = 0, + LowerLeft = 1, + }; + union { + BitField<0, 1, Mode> mode; + BitField<4, 1, u32> flip_y; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x3F); + struct IteratedBlendConstants { + u32 r; + u32 g; + u32 b; + INSERT_PADDING_BYTES_NOINIT(0x4); + }; + static_assert(sizeof(IteratedBlendConstants) == 0x10); + struct UserClip { + struct Enable { union { - BitField<0, 4, u32> stencil; - BitField<4, 4, u32> unknown; - BitField<8, 4, u32> scissor; - BitField<12, 4, u32> viewport; - } clear_flags; - - INSERT_PADDING_WORDS_NOINIT(0x10); - - u32 fill_rectangle; - - INSERT_PADDING_WORDS_NOINIT(0x2); - - u32 conservative_raster_enable; - - INSERT_PADDING_WORDS_NOINIT(0x5); - - std::array<VertexAttribute, NumVertexAttributes> vertex_attrib_format; + u32 raw; + BitField<0, 1, u32> plane0; + BitField<1, 1, u32> plane1; + BitField<2, 1, u32> plane2; + BitField<3, 1, u32> plane3; + BitField<4, 1, u32> plane4; + BitField<5, 1, u32> plane5; + BitField<6, 1, u32> plane6; + BitField<7, 1, u32> plane7; + }; - std::array<MsaaSampleLocation, 4> multisample_sample_locations; + bool AnyEnabled() const { + return plane0 || plane1 || plane2 || plane3 || plane4 || plane5 || plane6 || + plane7; + } + }; - INSERT_PADDING_WORDS_NOINIT(0x2); + struct Op { + enum class ClipOrCull : u32 { + Clip = 0, + Cull = 1, + }; union { - BitField<0, 1, u32> enable; - BitField<4, 3, u32> target; - } multisample_coverage_to_color; - - INSERT_PADDING_WORDS_NOINIT(0x8); - - struct { - union { - BitField<0, 4, u32> count; - BitField<4, 3, u32> map_0; - BitField<7, 3, u32> map_1; - BitField<10, 3, u32> map_2; - BitField<13, 3, u32> map_3; - BitField<16, 3, u32> map_4; - BitField<19, 3, u32> map_5; - BitField<22, 3, u32> map_6; - BitField<25, 3, u32> map_7; - }; - - u32 Map(std::size_t index) const { - const std::array<u32, NumRenderTargets> maps{map_0, map_1, map_2, map_3, - map_4, map_5, map_6, map_7}; - ASSERT(index < maps.size()); - return maps[index]; - } - } rt_control; - - INSERT_PADDING_WORDS_NOINIT(0x2); - - u32 zeta_width; - u32 zeta_height; - union { - BitField<0, 16, u32> zeta_depth; - BitField<16, 1, u32> zeta_volume; + u32 raw; + BitField<0, 1, ClipOrCull> plane0; + BitField<4, 1, ClipOrCull> plane1; + BitField<8, 1, ClipOrCull> plane2; + BitField<12, 1, ClipOrCull> plane3; + BitField<16, 1, ClipOrCull> plane4; + BitField<20, 1, ClipOrCull> plane5; + BitField<24, 1, ClipOrCull> plane6; + BitField<28, 1, ClipOrCull> plane7; }; + }; + }; - SamplerIndex sampler_index; + struct AntiAliasAlphaControl { + union { + BitField<0, 1, u32> alpha_to_coverage; + BitField<4, 1, u32> alpha_to_one; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x2); + struct RenderEnable { + enum class Override : u32 { + UseRenderEnable = 0, + AlwaysRender = 1, + NeverRender = 2, + }; - std::array<u32, 8> gp_passthrough_mask; + enum class Mode : u32 { + False = 0, + True = 1, + Conditional = 2, + IfEqual = 3, + IfNotEqual = 4, + }; - INSERT_PADDING_WORDS_NOINIT(0x1B); + u32 address_high; + u32 address_low; + Mode mode; - u32 depth_test_enable; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - INSERT_PADDING_WORDS_NOINIT(0x5); + struct TexSampler { + u32 address_high; + u32 address_low; + u32 limit; - u32 independent_blend_enable; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - u32 depth_write_enabled; + struct TexHeader { + u32 address_high; + u32 address_low; + u32 limit; - u32 alpha_test_enabled; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - INSERT_PADDING_WORDS_NOINIT(0x6); + enum class ZCullRegionFormat : u32 { + Z_4x4 = 0, + ZS_4x4 = 1, + Z_4x2 = 2, + Z_2x4 = 3, + Z_16x8_4x4 = 4, + Z_8x8_4x2 = 5, + Z_8x8_2x4 = 6, + Z_16x16_4x8 = 7, + Z_4x8_2x2 = 8, + ZS_16x8_4x2 = 9, + ZS_16x8_2x4 = 10, + ZS_8x8_2x2 = 11, + Z_4x8_1x1 = 12, + }; - u32 d3d_cull_mode; + struct RtLayer { + enum class Control { + LayerSelectsLayer = 0, + GeometryShaderSelectsLayer = 1, + }; - ComparisonOp depth_test_func; - float alpha_test_ref; - ComparisonOp alpha_test_func; - u32 draw_tfb_stride; - struct { - float r; - float g; - float b; - float a; - } blend_color; + union { + BitField<0, 16, u32> layer; + BitField<16, 1, u32> control; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x4); + struct InlineIndex2x16 { + union { + BitField<0, 31, u32> count; + BitField<31, 1, u32> start_odd; + }; + union { + BitField<0, 16, u32> even; + BitField<16, 16, u32> odd; + }; + }; - struct { - u32 separate_alpha; - Blend::Equation equation_rgb; - Blend::Factor factor_source_rgb; - Blend::Factor factor_dest_rgb; - Blend::Equation equation_a; - Blend::Factor factor_source_a; - INSERT_PADDING_WORDS_NOINIT(1); - Blend::Factor factor_dest_a; + struct VertexGlobalBaseOffset { + u32 address_high; + u32 address_low; - u32 enable_common; - u32 enable[NumRenderTargets]; - } blend; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - u32 stencil_enable; - StencilOp stencil_front_op_fail; - StencilOp stencil_front_op_zfail; - StencilOp stencil_front_op_zpass; - ComparisonOp stencil_front_func_func; - s32 stencil_front_func_ref; - u32 stencil_front_func_mask; - u32 stencil_front_mask; + struct ZCullRegionPixelOffset { + u32 width; + u32 height; + }; - INSERT_PADDING_WORDS_NOINIT(0x2); + struct PointSprite { + enum class RMode : u32 { + Zero = 0, + FromR = 1, + FromS = 2, + }; + enum class Origin : u32 { + Bottom = 0, + Top = 1, + }; + enum class Texture : u32 { + Passthrough = 0, + Generate = 1, + }; + + union { + BitField<0, 2, RMode> rmode; + BitField<2, 1, Origin> origin; + BitField<3, 1, Texture> texture0; + BitField<4, 1, Texture> texture1; + BitField<5, 1, Texture> texture2; + BitField<6, 1, Texture> texture3; + BitField<7, 1, Texture> texture4; + BitField<8, 1, Texture> texture5; + BitField<9, 1, Texture> texture6; + BitField<10, 1, Texture> texture7; + BitField<11, 1, Texture> texture8; + BitField<12, 1, Texture> texture9; + }; + }; - u32 frag_color_clamp; + struct ProgramRegion { + u32 address_high; + u32 address_low; - union { - BitField<0, 1, u32> y_negate; - BitField<4, 1, u32> triangle_rast_flip; - } screen_y_control; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - float line_width_smooth; - float line_width_aliased; + struct DefaultAttributes { + enum class Diffuse : u32 { + Vector_0001 = 0, + Vector_1111 = 1, + }; + enum class Specular : u32 { + Vector_0000 = 0, + Vector_0001 = 1, + }; + enum class Vector : u32 { + Vector_0000 = 0, + Vector_0001 = 1, + }; + enum class FixedFncTexture : u32 { + Vector_0000 = 0, + Vector_0001 = 1, + }; + enum class DX9Color0 : u32 { + Vector_0000 = 0, + Vector_1111 = 1, + }; + enum class DX9Color1To15 : u32 { + Vector_0000 = 0, + Vector_0001 = 1, + }; - INSERT_PADDING_WORDS_NOINIT(0x1B); + union { + BitField<0, 1, Diffuse> color_front_diffuse; + BitField<1, 1, Specular> color_front_specular; + BitField<2, 1, Vector> generic_vector; + BitField<3, 1, FixedFncTexture> fixed_fnc_texture; + BitField<4, 1, DX9Color0> dx9_color0; + BitField<5, 1, DX9Color1To15> dx9_color1_to_15; + }; + }; - u32 invalidate_sampler_cache_no_wfi; - u32 invalidate_texture_header_cache_no_wfi; + struct Draw { + enum class PrimitiveId : u32 { + First = 0, + Unchanged = 1, + }; + enum class InstanceId : u32 { + First = 0, + Subsequent = 1, + Unchanged = 2, + }; + enum class SplitMode : u32 { + NormalBeginNormal = 0, + NormalBeginOpen = 1, + OpenBeginOpen = 2, + OpenBeginNormal = 3, + }; - INSERT_PADDING_WORDS_NOINIT(0x2); + u32 end; + union { + u32 begin; + BitField<0, 16, PrimitiveTopology> topology; + BitField<24, 1, PrimitiveId> primitive_id; + BitField<26, 2, InstanceId> instance_id; + BitField<29, 2, SplitMode> split_mode; + }; + }; - u32 vb_element_base; - u32 vb_base_instance; + struct VertexIdCopy { + union { + BitField<0, 1, u32> enable; + BitField<4, 8, u32> attribute_slot; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x35); + struct ShaderBasedCull { + union { + BitField<1, 1, u32> batch_cull_enable; + BitField<0, 1, u32> before_fetch_enable; + }; + }; - u32 clip_distance_enabled; + struct ClassVersion { + union { + BitField<0, 16, u32> current; + BitField<16, 16, u32> oldest_supported; + }; + }; - u32 samplecnt_enable; + struct PrimitiveRestart { + u32 enabled; + u32 index; + }; - float point_size; + struct OutputVertexId { + union { + BitField<12, 1, u32> uses_array_start; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x1); + enum class PointCenterMode : u32 { + GL = 0, + D3D = 1, + }; - u32 point_sprite_enable; + enum class LineSmoothParams : u32 { + Falloff_1_00 = 0, + Falloff_1_33 = 1, + Falloff_1_66 = 2, + }; - INSERT_PADDING_WORDS_NOINIT(0x3); + struct LineSmoothEdgeTable { + union { + BitField<0, 8, u32> v0; + BitField<8, 8, u32> v1; + BitField<16, 8, u32> v2; + BitField<24, 8, u32> v3; + }; + }; - CounterReset counter_reset; + struct LineStippleParams { + union { + BitField<0, 8, u32> factor; + BitField<8, 16, u32> pattern; + }; + }; - u32 multisample_enable; + enum class ProvokingVertex : u32 { + First = 0, + Last = 1, + }; - u32 zeta_enable; + struct ShaderControl { + enum class Partial : u32 { + Zero = 0, + Infinity = 1, + }; + enum class FP32NanBehavior : u32 { + Legacy = 0, + FP64Compatible = 1, + }; + enum class FP32F2INanBehavior : u32 { + PassZero = 0, + PassIndefinite = 1, + }; - union { - BitField<0, 1, u32> alpha_to_coverage; - BitField<4, 1, u32> alpha_to_one; - } multisample_control; + union { + BitField<0, 1, Partial> default_partial; + BitField<1, 1, FP32NanBehavior> fp32_nan_behavior; + BitField<2, 1, FP32F2INanBehavior> fp32_f2i_nan_behavior; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x4); + struct SphVersion { + union { + BitField<0, 16, u32> current; + BitField<16, 16, u32> oldest_supported; + }; + }; - struct { - u32 address_high; - u32 address_low; - ConditionMode mode; + struct AlphaToCoverageOverride { + union { + BitField<0, 1, u32> qualify_by_anti_alias_enable; + BitField<1, 1, u32> qualify_by_ps_sample_mask_enable; + }; + }; - GPUVAddr Address() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | - address_low); - } - } condition; + struct AamVersion { + union { + BitField<0, 16, u32> current; + BitField<16, 16, u32> oldest_supported; + }; + }; - struct { - u32 address_high; - u32 address_low; - u32 limit; + struct IndexBuffer { + u32 start_addr_high; + u32 start_addr_low; + u32 limit_addr_high; + u32 limit_addr_low; + IndexFormat format; + u32 first; + u32 count; + + unsigned FormatSizeInBytes() const { + switch (format) { + case IndexFormat::UnsignedByte: + return 1; + case IndexFormat::UnsignedShort: + return 2; + case IndexFormat::UnsignedInt: + return 4; + } + ASSERT(false); + return 1; + } - GPUVAddr Address() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | - address_low); - } - } tsc; + GPUVAddr StartAddress() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(start_addr_high) << 32) | + start_addr_low); + } - INSERT_PADDING_WORDS_NOINIT(0x1); + GPUVAddr EndAddress() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(limit_addr_high) << 32) | + limit_addr_low); + } - float polygon_offset_factor; + /// Adjust the index buffer offset so it points to the first desired index. + GPUVAddr IndexStart() const { + return StartAddress() + + static_cast<size_t>(first) * static_cast<size_t>(FormatSizeInBytes()); + } + }; - u32 line_smooth_enable; + struct IndexBufferSmall { + union { + BitField<0, 16, u32> first; + BitField<16, 12, u32> count; + BitField<28, 4, PrimitiveTopology> topology; + }; + }; - struct { - u32 address_high; - u32 address_low; - u32 limit; + struct VertexStreamInstances { + std::array<u32, NumVertexArrays> is_instanced; - GPUVAddr Address() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | - address_low); - } - } tic; + /// Returns whether the vertex array specified by index is supposed to be + /// accessed per instance or not. + bool IsInstancingEnabled(std::size_t index) const { + return is_instanced[index]; + } + }; - INSERT_PADDING_WORDS_NOINIT(0x5); + struct AttributePointSize { + union { + BitField<0, 1, u32> enabled; + BitField<4, 8, u32> slot; + }; + }; - u32 stencil_two_side_enable; - StencilOp stencil_back_op_fail; - StencilOp stencil_back_op_zfail; - StencilOp stencil_back_op_zpass; - ComparisonOp stencil_back_func_func; + struct ViewportClipControl { + enum class GeometryGuardband : u32 { + Scale256 = 0, + Scale1 = 1, + }; + enum class GeometryClip : u32 { + WZero = 0, + Passthrough = 1, + FrustumXY = 2, + FrustumXYZ = 3, + WZeroNoZCull = 4, + FrustumZ = 5, + WZeroTriFillOrClip = 6, + }; + enum class GeometryGuardbandZ : u32 { + SameAsXY = 0, + Scale256 = 1, + Scale1 = 2, + }; - INSERT_PADDING_WORDS_NOINIT(0x4); + union { + BitField<0, 1, u32> depth_0_to_1; + BitField<3, 1, u32> pixel_min_z; + BitField<4, 1, u32> pixel_max_z; + BitField<7, 1, GeometryGuardband> geometry_guardband; + BitField<11, 3, GeometryClip> geometry_clip; + BitField<1, 2, GeometryGuardbandZ> geometry_guardband_z; + }; + }; - u32 framebuffer_srgb; + enum class PrimitiveTopologyControl : u32 { + UseInBeginMethods = 0, + UseSeparateState = 1, + }; - float polygon_offset_units; + struct WindowClip { + enum class Type : u32 { + Inclusive = 0, + Exclusive = 1, + ClipAll = 2, + }; - INSERT_PADDING_WORDS_NOINIT(0x4); + u32 enable; + Type type; + }; - Tegra::Texture::MsaaMode multisample_mode; + enum class InvalidateZCull : u32 { + Invalidate = 0, + }; - INSERT_PADDING_WORDS_NOINIT(0xC); + struct ZCull { + union { + BitField<0, 1, u32> z_enable; + BitField<1, 1, u32> stencil_enable; + }; + union { + BitField<0, 1, u32> z_min_enbounded; + BitField<1, 1, u32> z_max_unbounded; + }; + }; - union { - BitField<2, 1, u32> coord_origin; - BitField<3, 10, u32> enable; - } point_coord_replace; - - struct { - u32 code_address_high; - u32 code_address_low; - - GPUVAddr CodeAddress() const { - return static_cast<GPUVAddr>( - (static_cast<GPUVAddr>(code_address_high) << 32) | code_address_low); - } - } code_address; - INSERT_PADDING_WORDS_NOINIT(1); - - struct { - u32 vertex_end_gl; - union { - u32 vertex_begin_gl; - BitField<0, 16, PrimitiveTopology> topology; - BitField<26, 1, u32> instance_next; - BitField<27, 1, u32> instance_cont; - }; - } draw; - - INSERT_PADDING_WORDS_NOINIT(0xA); - - struct { - u32 enabled; - u32 index; - } primitive_restart; - - INSERT_PADDING_WORDS_NOINIT(0xE); - - u32 provoking_vertex_last; - - INSERT_PADDING_WORDS_NOINIT(0x50); - - struct { - u32 start_addr_high; - u32 start_addr_low; - u32 end_addr_high; - u32 end_addr_low; - IndexFormat format; - u32 first; - u32 count; - - unsigned FormatSizeInBytes() const { - switch (format) { - case IndexFormat::UnsignedByte: - return 1; - case IndexFormat::UnsignedShort: - return 2; - case IndexFormat::UnsignedInt: - return 4; - } - ASSERT(false); - return 1; - } - - GPUVAddr StartAddress() const { - return static_cast<GPUVAddr>( - (static_cast<GPUVAddr>(start_addr_high) << 32) | start_addr_low); - } - - GPUVAddr EndAddress() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(end_addr_high) << 32) | - end_addr_low); - } - - /// Adjust the index buffer offset so it points to the first desired index. - GPUVAddr IndexStart() const { - return StartAddress() + static_cast<size_t>(first) * - static_cast<size_t>(FormatSizeInBytes()); - } - } index_array; + struct LogicOp { + enum class Op : u32 { + Clear = 0x1500, + And = 0x1501, + AndReverse = 0x1502, + Copy = 0x1503, + AndInverted = 0x1504, + NoOp = 0x1505, + Xor = 0x1506, + Or = 0x1507, + Nor = 0x1508, + Equiv = 0x1509, + Invert = 0x150A, + OrReverse = 0x150B, + CopyInverted = 0x150C, + OrInverted = 0x150D, + Nand = 0x150E, + Set = 0x150F, + }; - union { - BitField<0, 16, u32> first; - BitField<16, 16, u32> count; - } small_index; + u32 enable; + Op op; + }; - union { - BitField<0, 16, u32> first; - BitField<16, 16, u32> count; - } small_index_2; + struct ClearSurface { + union { + u32 raw; + BitField<0, 1, u32> Z; + BitField<1, 1, u32> S; + BitField<2, 1, u32> R; + BitField<3, 1, u32> G; + BitField<4, 1, u32> B; + BitField<5, 1, u32> A; + BitField<6, 4, u32> RT; + BitField<10, 16, u32> layer; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x5); + struct ReportSemaphore { + struct Compare { + u32 initial_sequence; + u32 initial_mode; + u32 unknown1; + u32 unknown2; + u32 current_sequence; + u32 current_mode; + }; - INSERT_PADDING_WORDS_NOINIT(0x1F); + enum class Operation : u32 { + Release = 0, + Acquire = 1, + ReportOnly = 2, + Trap = 3, + }; - float polygon_offset_clamp; + enum class Release : u32 { + AfterAllPreceedingReads = 0, + AfterAllPreceedingWrites = 1, + }; - struct { - u32 is_instanced[NumVertexArrays]; + enum class Acquire : u32 { + BeforeAnyFollowingWrites = 0, + BeforeAnyFollowingReads = 1, + }; - /// Returns whether the vertex array specified by index is supposed to be - /// accessed per instance or not. - bool IsInstancingEnabled(std::size_t index) const { - return is_instanced[index]; - } - } instanced_arrays; + enum class Location : u32 { + None = 0, + VertexFetch = 1, + VertexShader = 2, + VPC = 4, + StreamingOutput = 5, + GeometryShader = 6, + ZCull = 7, + TessellationInit = 8, + TessellationShader = 9, + PixelShader = 10, + DepthTest = 12, + All = 15, + }; - INSERT_PADDING_WORDS_NOINIT(0x4); + enum class Comparison : u32 { + NotEqual = 0, + GreaterOrEqual = 1, + }; - union { - BitField<0, 1, u32> enable; - BitField<4, 8, u32> unk4; - } vp_point_size; + enum class Report : u32 { + Payload = 0, // "None" in docs, but confirmed via hardware to return the payload + VerticesGenerated = 1, + ZPassPixelCount = 2, + PrimitivesGenerated = 3, + AlphaBetaClocks = 4, + VertexShaderInvocations = 5, + StreamingPrimitivesNeededMinusSucceeded = 6, + GeometryShaderInvocations = 7, + GeometryShaderPrimitivesGenerated = 9, + ZCullStats0 = 10, + StreamingPrimitivesSucceeded = 11, + ZCullStats1 = 12, + StreamingPrimitivesNeeded = 13, + ZCullStats2 = 14, + ClipperInvocations = 15, + ZCullStats3 = 16, + ClipperPrimitivesGenerated = 17, + VtgPrimitivesOut = 18, + PixelShaderInvocations = 19, + ZPassPixelCount64 = 21, + IEEECleanColorTarget = 24, + IEEECleanZetaTarget = 25, + StreamingByteCount = 26, + TessellationInitInvocations = 27, + BoundingRectangle = 28, + TessellationShaderInvocations = 29, + TotalStreamingPrimitivesNeededMinusSucceeded = 30, + TessellationShaderPrimitivesGenerated = 31, + }; - INSERT_PADDING_WORDS_NOINIT(1); + u32 address_high; + u32 address_low; + u32 payload; + union { + u32 raw; + BitField<0, 2, Operation> operation; + BitField<4, 1, Release> release; + BitField<8, 1, Acquire> acquire; + BitField<12, 4, Location> location; + BitField<16, 1, Comparison> comparison; + BitField<20, 1, u32> awaken_enable; + BitField<23, 5, Report> report; + BitField<28, 1, u32> short_query; + BitField<5, 3, u32> sub_report; + BitField<21, 1, u32> dword_number; + BitField<2, 1, u32> disable_flush; + BitField<3, 1, u32> reduction_enable; + BitField<9, 3, ReductionOp> reduction_op; + BitField<17, 2, u32> format_signed; + } query; - u32 cull_test_enabled; - FrontFace front_face; - CullFace cull_face; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - u32 pixel_center_integer; + struct VertexStream { + union { + BitField<0, 12, u32> stride; + BitField<12, 1, u32> enable; + }; + u32 address_high; + u32 address_low; + u32 frequency; - INSERT_PADDING_WORDS_NOINIT(0x1); + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } - u32 viewport_transform_enabled; + bool IsEnabled() const { + return enable != 0 && Address() != 0; + } + }; + static_assert(sizeof(VertexStream) == 0x10); - INSERT_PADDING_WORDS_NOINIT(0x3); + struct VertexStreamLimit { + u32 address_high; + u32 address_low; - union { - BitField<0, 1, u32> depth_range_0_1; - BitField<3, 1, u32> depth_clamp_near; - BitField<4, 1, u32> depth_clamp_far; - BitField<11, 1, u32> depth_clamp_disabled; - } view_volume_clip_control; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; + static_assert(sizeof(VertexStreamLimit) == 0x8); - INSERT_PADDING_WORDS_NOINIT(0xC); + enum class ShaderType : u32 { + VertexA = 0, + VertexB = 1, + TessellationInit = 2, + Tessellation = 3, + Geometry = 4, + Pixel = 5, + }; - PrimitiveTopologyOverride topology_override; + struct Pipeline { + union { + BitField<0, 1, u32> enable; + BitField<4, 4, ShaderType> program; + }; + u32 offset; + u32 reservedA; + u32 register_count; + u32 binding_group; + std::array<u32, 4> reserved; + INSERT_PADDING_BYTES_NOINIT(0x1C); + }; + static_assert(sizeof(Pipeline) == 0x40); - INSERT_PADDING_WORDS_NOINIT(0x12); + bool IsShaderConfigEnabled(std::size_t index) const { + // The VertexB is always enabled. + if (index == static_cast<std::size_t>(ShaderType::VertexB)) { + return true; + } + return pipelines[index].enable != 0; + } - u32 depth_bounds_enable; + bool IsShaderConfigEnabled(ShaderType type) const { + return IsShaderConfigEnabled(static_cast<std::size_t>(type)); + } - INSERT_PADDING_WORDS_NOINIT(1); + struct ConstantBuffer { + u32 size; + u32 address_high; + u32 address_low; + u32 offset; + std::array<u32, NumCBData> buffer; - struct { - u32 enable; - LogicOperation operation; - } logic_op; + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); + } + }; - INSERT_PADDING_WORDS_NOINIT(0x1); + struct BindGroup { + std::array<u32, 4> reserved; + union { + u32 raw_config; + BitField<0, 1, u32> valid; + BitField<4, 5, u32> shader_slot; + }; + INSERT_PADDING_BYTES_NOINIT(0xC); + }; + static_assert(sizeof(BindGroup) == 0x20); + struct StreamOutLayout { + union { + BitField<0, 8, u32> attribute0; + BitField<8, 8, u32> attribute1; + BitField<16, 8, u32> attribute2; + BitField<24, 8, u32> attribute3; + }; + }; + + struct ShaderPerformance { + struct ControlA { union { - u32 raw; - BitField<0, 1, u32> Z; - BitField<1, 1, u32> S; - BitField<2, 1, u32> R; - BitField<3, 1, u32> G; - BitField<4, 1, u32> B; - BitField<5, 1, u32> A; - BitField<6, 4, u32> RT; - BitField<10, 11, u32> layer; - } clear_buffers; - INSERT_PADDING_WORDS_NOINIT(0xB); - std::array<ColorMask, NumRenderTargets> color_mask; - INSERT_PADDING_WORDS_NOINIT(0x38); - - struct { - u32 query_address_high; - u32 query_address_low; - u32 query_sequence; - union { - u32 raw; - BitField<0, 2, QueryOperation> operation; - BitField<4, 1, u32> fence; - BitField<12, 4, QueryUnit> unit; - BitField<16, 1, QuerySyncCondition> sync_cond; - BitField<23, 5, QuerySelect> select; - BitField<28, 1, u32> short_query; - } query_get; - - GPUVAddr QueryAddress() const { - return static_cast<GPUVAddr>( - (static_cast<GPUVAddr>(query_address_high) << 32) | query_address_low); - } - } query; - - INSERT_PADDING_WORDS_NOINIT(0x3C); - - struct { - union { - BitField<0, 12, u32> stride; - BitField<12, 1, u32> enable; - }; - u32 start_high; - u32 start_low; - u32 divisor; - - GPUVAddr StartAddress() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(start_high) << 32) | - start_low); - } - - bool IsEnabled() const { - return enable != 0 && StartAddress() != 0; - } - - } vertex_array[NumVertexArrays]; - - Blend independent_blend[NumRenderTargets]; - - struct { - u32 limit_high; - u32 limit_low; - - GPUVAddr LimitAddress() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(limit_high) << 32) | - limit_low); - } - } vertex_array_limit[NumVertexArrays]; - - struct { - union { - BitField<0, 1, u32> enable; - BitField<4, 4, ShaderProgram> program; - }; - u32 offset; - INSERT_PADDING_WORDS_NOINIT(14); - } shader_config[MaxShaderProgram]; - - INSERT_PADDING_WORDS_NOINIT(0x60); - - u32 firmware[0x20]; - - struct { - u32 cb_size; - u32 cb_address_high; - u32 cb_address_low; - u32 cb_pos; - std::array<u32, NumCBData> cb_data; - - GPUVAddr BufferAddress() const { - return static_cast<GPUVAddr>( - (static_cast<GPUVAddr>(cb_address_high) << 32) | cb_address_low); - } - } const_buffer; - - INSERT_PADDING_WORDS_NOINIT(0x10); - - struct { - union { - u32 raw_config; - BitField<0, 1, u32> valid; - BitField<4, 5, u32> index; - }; - INSERT_PADDING_WORDS_NOINIT(7); - } cb_bind[MaxShaderStage]; - - INSERT_PADDING_WORDS_NOINIT(0x56); - - u32 tex_cb_index; - - INSERT_PADDING_WORDS_NOINIT(0x7D); - - std::array<std::array<u8, 128>, NumTransformFeedbackBuffers> tfb_varying_locs; - - INSERT_PADDING_WORDS_NOINIT(0x298); - - struct { - /// Compressed address of a buffer that holds information about bound SSBOs. - /// This address is usually bound to c0 in the shaders. - u32 buffer_address; - - GPUVAddr BufferAddress() const { - return static_cast<GPUVAddr>(buffer_address) << 8; - } - } ssbo_info; + BitField<0, 2, u32> event0; + BitField<2, 3, u32> bit0; + BitField<5, 2, u32> event1; + BitField<7, 3, u32> bit1; + BitField<10, 2, u32> event2; + BitField<12, 3, u32> bit2; + BitField<15, 2, u32> event3; + BitField<17, 3, u32> bit3; + BitField<20, 2, u32> event4; + BitField<22, 3, u32> bit4; + BitField<25, 2, u32> event5; + BitField<27, 3, u32> bit5; + BitField<30, 2, u32> spare; + }; + }; - INSERT_PADDING_WORDS_NOINIT(0x11); + struct ControlB { + union { + BitField<0, 1, u32> edge; + BitField<1, 2, u32> mode; + BitField<3, 1, u32> windowed; + BitField<4, 16, u32> func; + }; + }; - struct { - u32 address[MaxShaderStage]; - u32 size[MaxShaderStage]; - } tex_info_buffers; + std::array<u32, 8> values_upper; + std::array<u32, 8> values; + std::array<u32, 8> events; + std::array<ControlA, 8> control_a; + std::array<ControlB, 8> control_b; + u32 trap_control_mask; + u32 start_shader_mask; + u32 stop_shader_mask; + }; - INSERT_PADDING_WORDS_NOINIT(0xCC); + // clang-format off + union { + struct { + ID object_id; ///< 0x0000 + INSERT_PADDING_BYTES_NOINIT(0xFC); + u32 nop; ///< 0x0100 + Notify notify; ///< 0x0104 + u32 wait_for_idle; ///< 0x0110 + LoadMME load_mme; ///< 0x0114 + ShadowRamControl shadow_ram_control; ///< 0x0124 + PeerSemaphore peer; ///< 0x0128 + GlobalRender global_render; ///< 0x0130 + u32 go_idle; ///< 0x013C + u32 trigger; ///< 0x0140 + u32 trigger_wfi; ///< 0x0144 + INSERT_PADDING_BYTES_NOINIT(0x8); + u32 instrumentation_method_header; ///< 0x0150 + u32 instrumentation_method_data; ///< 0x0154 + INSERT_PADDING_BYTES_NOINIT(0x28); + Upload::Registers upload; ///< 0x0180 + LaunchDMA launch_dma; ///< 0x01B0 + u32 inline_data; ///< 0x01B4 + INSERT_PADDING_BYTES_NOINIT(0x24); + I2M i2m; ///< 0x01DC + u32 run_ds_now; ///< 0x0200 + OpportunisticEarlyZ opportunistic_early_z; ///< 0x0204 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 aliased_line_width_enabled; ///< 0x020C + u32 mandated_early_z; ///< 0x0210 + GeometryShaderDmFifo gs_dm_fifo; ///< 0x0214 + L2CacheControl l2_cache_control; ///< 0x0218 + InvalidateShaderCache invalidate_shader_cache; ///< 0x021C + INSERT_PADDING_BYTES_NOINIT(0xA8); + SyncInfo sync_info; ///< 0x02C8 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 prim_circular_buffer_throttle; ///< 0x02D0 + u32 flush_invalidate_rop_mini_cache; ///< 0x02D4 + SurfaceClipBlockId surface_clip_block_id; ///< 0x02D8 + u32 alpha_circular_buffer_size; ///< 0x02DC + DecompressSurface decompress_surface; ///< 0x02E0 + ZCullRopBypass zcull_rop_bypass; ///< 0x02E4 + ZCullSubregion zcull_subregion; ///< 0x02E8 + RasterBoundingBox raster_bounding_box; ///< 0x02EC + u32 peer_semaphore_release; ///< 0x02F0 + u32 iterated_blend_optimization; ///< 0x02F4 + ZCullSubregionAllocation zcull_subregion_allocation; ///< 0x02F8 + ZCullSubregionAlgorithm zcull_subregion_algorithm; ///< 0x02FC + PixelShaderOutputSampleMaskUsage ps_output_sample_mask_usage; ///< 0x0300 + u32 draw_zero_index; ///< 0x0304 + L1Configuration l1_configuration; ///< 0x0308 + u32 render_enable_control_load_const_buffer; ///< 0x030C + SPAVersion spa_version; ///< 0x0310 + u32 ieee_clean_update; ///< 0x0314 + SnapGrid snap_grid; ///< 0x0318 + Tessellation tessellation; ///< 0x0320 + SubTilingPerf sub_tiling_perf; ///< 0x0360 + ZCullSubregionReport zcull_subregion_report; ///< 0x036C + BalancedPrimitiveWorkload balanced_primitive_workload; ///< 0x0374 + u32 max_patches_per_batch; ///< 0x0378 + u32 rasterize_enable; ///< 0x037C + TransformFeedback transform_feedback; ///< 0x0380 + u32 raster_input; ///< 0x0740 + u32 transform_feedback_enabled; ///< 0x0744 + u32 primitive_restart_topology_change_enable; ///< 0x0748 + u32 alpha_fraction; ///< 0x074C + INSERT_PADDING_BYTES_NOINIT(0x4); + HybridAntiAliasControl hybrid_aa_control; ///< 0x0754 + INSERT_PADDING_BYTES_NOINIT(0x24); + ShaderLocalMemory shader_local_memory; ///< 0x077C + u32 color_zero_bandwidth_clear; ///< 0x07A4 + u32 z_zero_bandwidth_clear; ///< 0x07A8 + u32 isbe_save_restore_program_offset; ///< 0x07AC + INSERT_PADDING_BYTES_NOINIT(0x10); + ZCullRegion zcull_region; ///< 0x07C0 + ZetaReadOnly zeta_read_only; ///< 0x07F8 + INSERT_PADDING_BYTES_NOINIT(0x4); + std::array<RenderTargetConfig, NumRenderTargets> rt; ///< 0x0800 + std::array<ViewportTransform, NumViewports> viewport_transform; ///< 0x0A00 + std::array<Viewport, NumViewports> viewports; ///< 0x0C00 + std::array<Window, 8> windows; ///< 0x0D00 + std::array<ClipIdExtent, 4> clip_id_extent; ///< 0x0D40 + u32 max_geometry_instances_per_task; ///< 0x0D60 + VisibleCallLimit visible_call_limit; ///< 0x0D64 + StatisticsCounter statistics_count; ///< 0x0D68 + ClearRect clear_rect; ///< 0x0D6C + VertexBuffer vertex_buffer; ///< 0x0D74 + DepthMode depth_mode; ///< 0x0D7C + std::array<f32, 4> clear_color; ///< 0x0D80 + f32 clear_depth; ///< 0x0D90 + u32 shader_cache_icache_prefetch; ///< 0x0D94 + u32 force_transition_to_beta; ///< 0x0D98 + u32 reduce_colour_thresholds; ///< 0x0D9C + s32 clear_stencil; ///< 0x0DA0 + InvalidateShaderCacheNoWFI invalidate_shader_cache_no_wfi; ///< 0x0DA4 + ZCullSerialization zcull_serialization; ///< 0x0DA8 + PolygonMode polygon_mode_front; ///< 0x0DAC + PolygonMode polygon_mode_back; ///< 0x0DB0 + u32 polygon_smooth; ///< 0x0DB4 + u32 zeta_mark_clean_ieee; ///< 0x0DB8 + ZCullDirFormat zcull_dir_format; ///< 0x0DBC + u32 polygon_offset_point_enable; ///< 0x0DC0 + u32 polygon_offset_line_enable; ///< 0x0DC4 + u32 polygon_offset_fill_enable; ///< 0x0DC8 + u32 patch_vertices; ///< 0x0DCC + IteratedBlend iterated_blend; ///< 0x0DD0 + ZCullCriterion zcull_criteria; ///< 0x0DD8 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 fragment_barrier; ///< 0x0DE0 + u32 sm_timeout; ///< 0x0DE4 + u32 primitive_restart_array; ///< 0x0DE8 + INSERT_PADDING_BYTES_NOINIT(0x4); + LoadIteratedBlend load_iterated_blend; ///< 0x0DF0 + u32 window_offset_x; ///< 0x0DF8 + u32 window_offset_y; ///< 0x0DFC + std::array<ScissorTest, NumViewports> scissor_test; ///< 0x0E00 + INSERT_PADDING_BYTES_NOINIT(0x10); + u32 select_texture_headers; ///< 0x0F10 + VPCPerf vpc_perf; ///< 0x0F14 + u32 pm_local_trigger; ///< 0x0F18 + u32 post_z_pixel_imask; ///< 0x0F1C + INSERT_PADDING_BYTES_NOINIT(0x20); + ConstantColorRendering const_color_rendering; ///< 0x0F40 + s32 stencil_back_ref; ///< 0x0F54 + u32 stencil_back_mask; ///< 0x0F58 + u32 stencil_back_func_mask; ///< 0x0F5C + INSERT_PADDING_BYTES_NOINIT(0x24); + VertexStreamSubstitute vertex_stream_substitute; ///< 0x0F84 + u32 line_mode_clip_generated_edge_do_not_draw; ///< 0x0F8C + u32 color_mask_common; ///< 0x0F90 + INSERT_PADDING_BYTES_NOINIT(0x4); + VTGWarpWatermarks vtg_warp_watermarks; ///< 0x0F98 + f32 depth_bounds[2]; ///< 0x0F9C + SampleMask::Target sample_mask_target; ///< 0x0FA4 + u32 color_target_mrt_enable; ///< 0x0FAC + NonMultisampledZ non_multisampled_z; ///< 0x0FB0 + TIRMode tir_mode; ///< 0x0FB4 + AntiAliasRaster anti_alias_raster; ///< 0x0FB8 + SampleMask::Pos sample_mask_pos; ///< 0x0FBC + SurfaceClipIDMemory surface_clip_id_memory; ///< 0x0FCC + TIRModulation tir_modulation; ///< 0x0FD4 + u32 blend_control_allow_float_pixel_kills; ///< 0x0FDC + Zeta zeta; ///< 0x0FE0 + SurfaceClip surface_clip; ///< 0x0FF4 + u32 tiled_cache_treat_heavy_as_light; ///< 0x0FFC + L2CacheVAFRequests l2_cache_vaf; ///< 0x1000 + ViewportMulticast viewport_multicast; ///< 0x1004 + u32 tessellation_cut_height; ///< 0x1008 + u32 max_gs_instances_per_task; ///< 0x100C + u32 max_gs_output_vertices_per_task; ///< 0x1010 + u32 reserved_sw_method0; ///< 0x1014 + u32 gs_output_cb_storage_multiplier; ///< 0x1018 + u32 beta_cb_storage_constant; ///< 0x101C + u32 ti_output_cb_storage_multiplier; ///< 0x1020 + u32 alpha_cb_storage_constraint; ///< 0x1024 + u32 reserved_sw_method1; ///< 0x1028 + u32 reserved_sw_method2; ///< 0x102C + std::array<TIRModulationCoeff, 5> tir_modulation_coeff; ///< 0x1030 + std::array<u32, 15> spare_nop; ///< 0x1044 + INSERT_PADDING_BYTES_NOINIT(0x30); + std::array<u32, 7> reserved_sw_method3_to_7; ///< 0x10B0 + ReduceColorThreshold reduce_color_thresholds_unorm8; ///< 0x10CC + std::array<u32, 4> reserved_sw_method10_to_13; ///< 0x10D0 + ReduceColorThreshold reduce_color_thresholds_unorm10; ///< 0x10E0 + ReduceColorThreshold reduce_color_thresholds_unorm16; ///< 0x10E4 + ReduceColorThreshold reduce_color_thresholds_fp11; ///< 0x10E8 + ReduceColorThreshold reduce_color_thresholds_fp16; ///< 0x10EC + ReduceColorThreshold reduce_color_thresholds_srgb8; ///< 0x10F0 + u32 unbind_all_constant_buffers; ///< 0x10F4 + ClearControl clear_control; ///< 0x10F8 + L2CacheRopNonInterlockedReads l2_cache_rop_non_interlocked_reads; ///< 0x10FC + u32 reserved_sw_method14; ///< 0x1100 + u32 reserved_sw_method15; ///< 0x1104 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 no_operation_data_high; ///< 0x110C + u32 depth_bias_control; ///< 0x1110 + u32 pm_trigger_end; ///< 0x1114 + u32 vertex_id_base; ///< 0x1118 + u32 stencil_compression_enabled; ///< 0x111C + VertexOutputAttributeSkipMasks vertex_output_attribute_skip_masks; ///< 0x1120 + TIRControl tir_control; ///< 0x1130 + u32 mutable_method_treat_mutable_as_heavy; ///< 0x1134 + u32 post_ps_use_pre_ps_coverage; ///< 0x1138 + FillViaTriangleMode fill_via_triangle_mode; ///< 0x113C + u32 blend_per_format_snorm8_unorm16_snorm16_enabled; ///< 0x1140 + u32 flush_pending_writes_sm_gloal_store; ///< 0x1144 + INSERT_PADDING_BYTES_NOINIT(0x18); + std::array<VertexAttribute, NumVertexAttributes> vertex_attrib_format; ///< 0x1160 + std::array<MsaaSampleLocation, 4> multisample_sample_locations; ///< 0x11E0 + u32 offset_render_target_index_by_viewport_index; ///< 0x11F0 + u32 force_heavy_method_sync; ///< 0x11F4 + MultisampleCoverageToColor multisample_coverage_to_color; ///< 0x11F8 + DecompressZetaSurface decompress_zeta_surface; ///< 0x11FC + INSERT_PADDING_BYTES_NOINIT(0x8); + ZetaSparse zeta_sparse; ///< 0x1208 + u32 invalidate_sampler_cache; ///< 0x120C + u32 invalidate_texture_header_cache; ///< 0x1210 + VertexArray vertex_array_instance_first; ///< 0x1214 + VertexArray vertex_array_instance_subsequent; ///< 0x1218 + RtControl rt_control; ///< 0x121C + CompressionThresholdSamples compression_threshold_samples; ///< 0x1220 + PixelShaderInterlockControl ps_interlock_control; ///< 0x1224 + ZetaSize zeta_size; ///< 0x1228 + SamplerBinding sampler_binding; ///< 0x1234 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 draw_auto_byte_count; ///< 0x123C + std::array<u32, 8> post_vtg_shader_attrib_skip_mask; ///< 0x1240 + PsTicketDispenserValue ps_ticket_dispenser_value; ///< 0x1260 + INSERT_PADDING_BYTES_NOINIT(0x1C); + u32 circular_buffer_size; ///< 0x1280 + RegisterWatermarks vtg_register_watermarks; ///< 0x1284 + InvalidateTextureDataCacheNoWfi invalidate_texture_cache_no_wfi; ///< 0x1288 + INSERT_PADDING_BYTES_NOINIT(0x4); + L2CacheRopNonInterlockedReads l2_cache_rop_interlocked_reads; ///< 0x1290 + INSERT_PADDING_BYTES_NOINIT(0x10); + u32 primitive_restart_topology_change_index; ///< 0x12A4 + INSERT_PADDING_BYTES_NOINIT(0x20); + ZCullRegionEnable zcull_region_enable; ///< 0x12C8 + u32 depth_test_enable; ///< 0x12CC + FillMode fill_mode; ///< 0x12D0 + ShadeMode shade_mode; ///< 0x12D4 + L2CacheRopNonInterlockedReads l2_cache_rop_non_interlocked_writes; ///< 0x12D8 + L2CacheRopNonInterlockedReads l2_cache_rop_interlocked_writes; ///< 0x12DC + AlphaToCoverageDither alpha_to_coverage_dither; ///< 0x12E0 + u32 blend_per_target_enabled; ///< 0x12E4 + u32 depth_write_enabled; ///< 0x12E8 + u32 alpha_test_enabled; ///< 0x12EC + INSERT_PADDING_BYTES_NOINIT(0x10); + InlineIndex4x8Align inline_index_4x8_align; ///< 0x1300 + InlineIndex4x8Index inline_index_4x8_index; ///< 0x1304 + D3DCullMode d3d_cull_mode; ///< 0x1308 + ComparisonOp depth_test_func; ///< 0x130C + f32 alpha_test_ref; ///< 0x1310 + ComparisonOp alpha_test_func; ///< 0x1314 + u32 draw_auto_stride; ///< 0x1318 + BlendColor blend_color; ///< 0x131C + INSERT_PADDING_BYTES_NOINIT(0x4); + InvalidateCacheLines invalidate_sampler_cache_lines; ///< 0x1330 + InvalidateCacheLines invalidate_texture_header_cache_lines; ///< 0x1334 + InvalidateCacheLines invalidate_texture_data_cache_lines; ///< 0x1338 + Blend blend; ///< 0x133C + u32 stencil_enable; ///< 0x1380 + StencilOp stencil_front_op; ///< 0x1384 + s32 stencil_front_ref; ///< 0x1394 + s32 stencil_front_func_mask; ///< 0x1398 + s32 stencil_front_mask; ///< 0x139C + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 draw_auto_start_byte_count; ///< 0x13A4 + PsSaturate frag_color_clamp; ///< 0x13A8 + WindowOrigin window_origin; ///< 0x13AC + f32 line_width_smooth; ///< 0x13B0 + f32 line_width_aliased; ///< 0x13B4 + INSERT_PADDING_BYTES_NOINIT(0x60); + u32 line_override_multisample; ///< 0x1418 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 alpha_hysteresis_rounds; ///< 0x1420 + InvalidateCacheLines invalidate_sampler_cache_no_wfi; ///< 0x1424 + InvalidateCacheLines invalidate_texture_header_cache_no_wfi; ///< 0x1428 + INSERT_PADDING_BYTES_NOINIT(0x8); + u32 global_base_vertex_index; ///< 0x1434 + u32 global_base_instance_index; ///< 0x1438 + INSERT_PADDING_BYTES_NOINIT(0x14); + RegisterWatermarks ps_warp_watermarks; ///< 0x1450 + RegisterWatermarks ps_regster_watermarks; ///< 0x1454 + INSERT_PADDING_BYTES_NOINIT(0xC); + u32 store_zcull; ///< 0x1464 + INSERT_PADDING_BYTES_NOINIT(0x18); + std::array<IteratedBlendConstants, NumRenderTargets> + iterated_blend_constants; ///< 0x1480 + u32 load_zcull; ///< 0x1500 + u32 surface_clip_id_height; ///< 0x1504 + Window surface_clip_id_clear_rect; ///< 0x1508 + UserClip::Enable user_clip_enable; ///< 0x1510 + u32 zpass_pixel_count_enable; ///< 0x1514 + f32 point_size; ///< 0x1518 + u32 zcull_stats_enable; ///< 0x151C + u32 point_sprite_enable; ///< 0x1520 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 shader_exceptions_enable; ///< 0x1528 + INSERT_PADDING_BYTES_NOINIT(0x4); + ClearReport clear_report_value; ///< 0x1530 + u32 anti_alias_enable; ///< 0x1534 + u32 zeta_enable; ///< 0x1538 + AntiAliasAlphaControl anti_alias_alpha_control; ///< 0x153C + INSERT_PADDING_BYTES_NOINIT(0x10); + RenderEnable render_enable; ///< 0x1550 + TexSampler tex_sampler; ///< 0x155C + INSERT_PADDING_BYTES_NOINIT(0x4); + f32 slope_scale_depth_bias; ///< 0x156C + u32 line_anti_alias_enable; ///< 0x1570 + TexHeader tex_header; ///< 0x1574 + INSERT_PADDING_BYTES_NOINIT(0x10); + u32 active_zcull_region_id; ///< 0x1590 + u32 stencil_two_side_enable; ///< 0x1594 + StencilOp stencil_back_op; ///< 0x1598 + INSERT_PADDING_BYTES_NOINIT(0x10); + u32 framebuffer_srgb; ///< 0x15B8 + f32 depth_bias; ///< 0x15BC + INSERT_PADDING_BYTES_NOINIT(0x8); + ZCullRegionFormat zcull_region_format; ///< 0x15C8 + RtLayer rt_layer; ///< 0x15CC + Tegra::Texture::MsaaMode anti_alias_samples_mode; ///< 0x15D0 + INSERT_PADDING_BYTES_NOINIT(0x10); + u32 edge_flag; ///< 0x15E4 + u32 draw_inline_index; ///< 0x15E8 + InlineIndex2x16 inline_index_2x16; ///< 0x15EC + VertexGlobalBaseOffset vertex_global_base_offset; ///< 0x15F4 + ZCullRegionPixelOffset zcull_region_pixel_offset; ///< 0x15FC + PointSprite point_sprite; ///< 0x1604 + ProgramRegion program_region; ///< 0x1608 + DefaultAttributes default_attributes; ///< 0x1610 + Draw draw; ///< 0x1614 + VertexIdCopy vertex_id_copy; ///< 0x161C + u32 add_to_primitive_id; ///< 0x1620 + u32 load_to_primitive_id; ///< 0x1624 + INSERT_PADDING_BYTES_NOINIT(0x4); + ShaderBasedCull shader_based_cull; ///< 0x162C + INSERT_PADDING_BYTES_NOINIT(0x8); + ClassVersion class_version; ///< 0x1638 + INSERT_PADDING_BYTES_NOINIT(0x8); + PrimitiveRestart primitive_restart; ///< 0x1644 + OutputVertexId output_vertex_id; ///< 0x164C + INSERT_PADDING_BYTES_NOINIT(0x8); + u32 anti_alias_point_enable; ///< 0x1658 + PointCenterMode point_center_mode; ///< 0x165C + INSERT_PADDING_BYTES_NOINIT(0x8); + LineSmoothParams line_smooth_params; ///< 0x1668 + u32 line_stipple_enable; ///< 0x166C + std::array<LineSmoothEdgeTable, 4> line_smooth_edge_table; ///< 0x1670 + LineStippleParams line_stipple_params; ///< 0x1680 + ProvokingVertex provoking_vertex; ///< 0x1684 + u32 two_sided_light_enabled; ///< 0x1688 + u32 polygon_stipple_enabled; ///< 0x168C + ShaderControl shader_control; ///< 0x1690 + INSERT_PADDING_BYTES_NOINIT(0xC); + ClassVersion class_version_check; ///< 0x16A0 + SphVersion sph_version; ///< 0x16A4 + SphVersion sph_version_check; ///< 0x16A8 + INSERT_PADDING_BYTES_NOINIT(0x8); + AlphaToCoverageOverride alpha_to_coverage_override; ///< 0x16B4 + INSERT_PADDING_BYTES_NOINIT(0x48); + std::array<u32, 32> polygon_stipple_pattern; ///< 0x1700 + INSERT_PADDING_BYTES_NOINIT(0x10); + AamVersion aam_version; ///< 0x1790 + AamVersion aam_version_check; ///< 0x1794 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 zeta_layer_offset; ///< 0x179C + INSERT_PADDING_BYTES_NOINIT(0x28); + IndexBuffer index_buffer; ///< 0x17C8 + IndexBufferSmall index_buffer32_first; ///< 0x17E4 + IndexBufferSmall index_buffer16_first; ///< 0x17E8 + IndexBufferSmall index_buffer8_first; ///< 0x17EC + IndexBufferSmall index_buffer32_subsequent; ///< 0x17F0 + IndexBufferSmall index_buffer16_subsequent; ///< 0x17F4 + IndexBufferSmall index_buffer8_subsequent; ///< 0x17F8 + INSERT_PADDING_BYTES_NOINIT(0x80); + f32 depth_bias_clamp; ///< 0x187C + VertexStreamInstances vertex_stream_instances; ///< 0x1880 + INSERT_PADDING_BYTES_NOINIT(0x10); + AttributePointSize point_size_attribute; ///< 0x1910 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 gl_cull_test_enabled; ///< 0x1918 + FrontFace gl_front_face; ///< 0x191C + CullFace gl_cull_face; ///< 0x1920 + Viewport::PixelCenter viewport_pixel_center; ///< 0x1924 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 viewport_scale_offset_enbled; ///< 0x192C + INSERT_PADDING_BYTES_NOINIT(0xC); + ViewportClipControl viewport_clip_control; ///< 0x193C + UserClip::Op user_clip_op; ///< 0x1940 + RenderEnable::Override render_enable_override; ///< 0x1944 + PrimitiveTopologyControl primitive_topology_control; ///< 0x1948 + WindowClip window_clip_enable; ///< 0x194C + INSERT_PADDING_BYTES_NOINIT(0x4); + InvalidateZCull invalidate_zcull; ///< 0x1958 + INSERT_PADDING_BYTES_NOINIT(0xC); + ZCull zcull; ///< 0x1968 + PrimitiveTopologyOverride topology_override; ///< 0x1970 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 zcull_sync; ///< 0x1978 + u32 clip_id_test_enable; ///< 0x197C + u32 surface_clip_id_width; ///< 0x1980 + u32 clip_id; ///< 0x1984 + INSERT_PADDING_BYTES_NOINIT(0x34); + u32 depth_bounds_enable; ///< 0x19BC + u32 blend_float_zero_times_anything_is_zero; ///< 0x19C0 + LogicOp logic_op; ///< 0x19C4 + u32 z_compression_enable; ///< 0x19CC + ClearSurface clear_surface; ///< 0x19D0 + u32 clear_clip_id_surface; ///< 0x19D4 + INSERT_PADDING_BYTES_NOINIT(0x8); + std::array<u32, NumRenderTargets> color_compression_enable; ///< 0x19E0 + std::array<ColorMask, NumRenderTargets> color_mask; ///< 0x1A00 + INSERT_PADDING_BYTES_NOINIT(0xC); + u32 pipe_nop; ///< 0x1A2C + std::array<u32, 4> spare; ///< 0x1A30 + INSERT_PADDING_BYTES_NOINIT(0xC0); + ReportSemaphore report_semaphore; ///< 0x1B00 + INSERT_PADDING_BYTES_NOINIT(0xF0); + std::array<VertexStream, NumVertexArrays> vertex_streams; ///< 0x1C00 + BlendPerTarget blend_per_target[NumRenderTargets]; ///< 0x1E00 + std::array<VertexStreamLimit, NumVertexArrays> vertex_stream_limits; ///< 0x1F00 + std::array<Pipeline, MaxShaderProgram> pipelines; ///< 0x2000 + INSERT_PADDING_BYTES_NOINIT(0x180); + u32 falcon[32]; ///< 0x2300 + ConstantBuffer const_buffer; ///< 0x2380 + INSERT_PADDING_BYTES_NOINIT(0x30); + BindGroup bind_groups[MaxShaderStage]; ///< 0x2400 + INSERT_PADDING_BYTES_NOINIT(0x160); + u32 color_clamp_enable; ///< 0x2600 + INSERT_PADDING_BYTES_NOINIT(0x4); + u32 bindless_texture_const_buffer_slot; ///< 0x2608 + u32 trap_handler; ///< 0x260C + INSERT_PADDING_BYTES_NOINIT(0x1F0); + std::array<std::array<StreamOutLayout, 32>, NumTransformFeedbackBuffers> + stream_out_layout; ///< 0x2800 + INSERT_PADDING_BYTES_NOINIT(0x93C); + ShaderPerformance shader_performance; ///< 0x333C + INSERT_PADDING_BYTES_NOINIT(0x18); + std::array<u32, 0x100> shadow_scratch; ///< 0x3400 }; std::array<u32, NUM_REGS> reg_array; }; }; + // clang-format on Regs regs{}; @@ -1591,147 +3201,350 @@ private: }; #define ASSERT_REG_POSITION(field_name, position) \ - static_assert(offsetof(Maxwell3D::Regs, field_name) == position * 4, \ + static_assert(offsetof(Maxwell3D::Regs, field_name) == position, \ "Field " #field_name " has invalid position") -ASSERT_REG_POSITION(wait_for_idle, 0x44); -ASSERT_REG_POSITION(macros, 0x45); -ASSERT_REG_POSITION(shadow_ram_control, 0x49); -ASSERT_REG_POSITION(upload, 0x60); -ASSERT_REG_POSITION(exec_upload, 0x6C); -ASSERT_REG_POSITION(data_upload, 0x6D); -ASSERT_REG_POSITION(force_early_fragment_tests, 0x84); -ASSERT_REG_POSITION(sync_info, 0xB2); -ASSERT_REG_POSITION(tess_mode, 0xC8); -ASSERT_REG_POSITION(tess_level_outer, 0xC9); -ASSERT_REG_POSITION(tess_level_inner, 0xCD); -ASSERT_REG_POSITION(rasterize_enable, 0xDF); -ASSERT_REG_POSITION(tfb_bindings, 0xE0); -ASSERT_REG_POSITION(tfb_layouts, 0x1C0); -ASSERT_REG_POSITION(tfb_enabled, 0x1D1); -ASSERT_REG_POSITION(rt, 0x200); -ASSERT_REG_POSITION(viewport_transform, 0x280); -ASSERT_REG_POSITION(viewports, 0x300); -ASSERT_REG_POSITION(vertex_buffer, 0x35D); -ASSERT_REG_POSITION(depth_mode, 0x35F); -ASSERT_REG_POSITION(clear_color[0], 0x360); -ASSERT_REG_POSITION(clear_depth, 0x364); -ASSERT_REG_POSITION(clear_stencil, 0x368); -ASSERT_REG_POSITION(polygon_mode_front, 0x36B); -ASSERT_REG_POSITION(polygon_mode_back, 0x36C); -ASSERT_REG_POSITION(polygon_offset_point_enable, 0x370); -ASSERT_REG_POSITION(polygon_offset_line_enable, 0x371); -ASSERT_REG_POSITION(polygon_offset_fill_enable, 0x372); -ASSERT_REG_POSITION(patch_vertices, 0x373); -ASSERT_REG_POSITION(fragment_barrier, 0x378); -ASSERT_REG_POSITION(scissor_test, 0x380); -ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5); -ASSERT_REG_POSITION(stencil_back_mask, 0x3D6); -ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7); -ASSERT_REG_POSITION(invalidate_texture_data_cache, 0x3DD); -ASSERT_REG_POSITION(tiled_cache_barrier, 0x3DF); -ASSERT_REG_POSITION(color_mask_common, 0x3E4); -ASSERT_REG_POSITION(depth_bounds, 0x3E7); -ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB); -ASSERT_REG_POSITION(multisample_raster_enable, 0x3ED); -ASSERT_REG_POSITION(multisample_raster_samples, 0x3EE); -ASSERT_REG_POSITION(multisample_sample_mask, 0x3EF); -ASSERT_REG_POSITION(zeta, 0x3F8); -ASSERT_REG_POSITION(render_area, 0x3FD); -ASSERT_REG_POSITION(clear_flags, 0x43E); -ASSERT_REG_POSITION(fill_rectangle, 0x44F); -ASSERT_REG_POSITION(conservative_raster_enable, 0x452); -ASSERT_REG_POSITION(vertex_attrib_format, 0x458); -ASSERT_REG_POSITION(multisample_sample_locations, 0x478); -ASSERT_REG_POSITION(multisample_coverage_to_color, 0x47E); -ASSERT_REG_POSITION(rt_control, 0x487); -ASSERT_REG_POSITION(zeta_width, 0x48a); -ASSERT_REG_POSITION(zeta_height, 0x48b); -ASSERT_REG_POSITION(zeta_depth, 0x48c); -ASSERT_REG_POSITION(sampler_index, 0x48D); -ASSERT_REG_POSITION(gp_passthrough_mask, 0x490); -ASSERT_REG_POSITION(depth_test_enable, 0x4B3); -ASSERT_REG_POSITION(independent_blend_enable, 0x4B9); -ASSERT_REG_POSITION(depth_write_enabled, 0x4BA); -ASSERT_REG_POSITION(alpha_test_enabled, 0x4BB); -ASSERT_REG_POSITION(d3d_cull_mode, 0x4C2); -ASSERT_REG_POSITION(depth_test_func, 0x4C3); -ASSERT_REG_POSITION(alpha_test_ref, 0x4C4); -ASSERT_REG_POSITION(alpha_test_func, 0x4C5); -ASSERT_REG_POSITION(draw_tfb_stride, 0x4C6); -ASSERT_REG_POSITION(blend_color, 0x4C7); -ASSERT_REG_POSITION(blend, 0x4CF); -ASSERT_REG_POSITION(stencil_enable, 0x4E0); -ASSERT_REG_POSITION(stencil_front_op_fail, 0x4E1); -ASSERT_REG_POSITION(stencil_front_op_zfail, 0x4E2); -ASSERT_REG_POSITION(stencil_front_op_zpass, 0x4E3); -ASSERT_REG_POSITION(stencil_front_func_func, 0x4E4); -ASSERT_REG_POSITION(stencil_front_func_ref, 0x4E5); -ASSERT_REG_POSITION(stencil_front_func_mask, 0x4E6); -ASSERT_REG_POSITION(stencil_front_mask, 0x4E7); -ASSERT_REG_POSITION(frag_color_clamp, 0x4EA); -ASSERT_REG_POSITION(screen_y_control, 0x4EB); -ASSERT_REG_POSITION(line_width_smooth, 0x4EC); -ASSERT_REG_POSITION(line_width_aliased, 0x4ED); -ASSERT_REG_POSITION(invalidate_sampler_cache_no_wfi, 0x509); -ASSERT_REG_POSITION(invalidate_texture_header_cache_no_wfi, 0x50A); -ASSERT_REG_POSITION(vb_element_base, 0x50D); -ASSERT_REG_POSITION(vb_base_instance, 0x50E); -ASSERT_REG_POSITION(clip_distance_enabled, 0x544); -ASSERT_REG_POSITION(samplecnt_enable, 0x545); -ASSERT_REG_POSITION(point_size, 0x546); -ASSERT_REG_POSITION(point_sprite_enable, 0x548); -ASSERT_REG_POSITION(counter_reset, 0x54C); -ASSERT_REG_POSITION(multisample_enable, 0x54D); -ASSERT_REG_POSITION(zeta_enable, 0x54E); -ASSERT_REG_POSITION(multisample_control, 0x54F); -ASSERT_REG_POSITION(condition, 0x554); -ASSERT_REG_POSITION(tsc, 0x557); -ASSERT_REG_POSITION(polygon_offset_factor, 0x55B); -ASSERT_REG_POSITION(line_smooth_enable, 0x55C); -ASSERT_REG_POSITION(tic, 0x55D); -ASSERT_REG_POSITION(stencil_two_side_enable, 0x565); -ASSERT_REG_POSITION(stencil_back_op_fail, 0x566); -ASSERT_REG_POSITION(stencil_back_op_zfail, 0x567); -ASSERT_REG_POSITION(stencil_back_op_zpass, 0x568); -ASSERT_REG_POSITION(stencil_back_func_func, 0x569); -ASSERT_REG_POSITION(framebuffer_srgb, 0x56E); -ASSERT_REG_POSITION(polygon_offset_units, 0x56F); -ASSERT_REG_POSITION(multisample_mode, 0x574); -ASSERT_REG_POSITION(point_coord_replace, 0x581); -ASSERT_REG_POSITION(code_address, 0x582); -ASSERT_REG_POSITION(draw, 0x585); -ASSERT_REG_POSITION(primitive_restart, 0x591); -ASSERT_REG_POSITION(provoking_vertex_last, 0x5A1); -ASSERT_REG_POSITION(index_array, 0x5F2); -ASSERT_REG_POSITION(small_index, 0x5F9); -ASSERT_REG_POSITION(polygon_offset_clamp, 0x61F); -ASSERT_REG_POSITION(instanced_arrays, 0x620); -ASSERT_REG_POSITION(vp_point_size, 0x644); -ASSERT_REG_POSITION(cull_test_enabled, 0x646); -ASSERT_REG_POSITION(front_face, 0x647); -ASSERT_REG_POSITION(cull_face, 0x648); -ASSERT_REG_POSITION(pixel_center_integer, 0x649); -ASSERT_REG_POSITION(viewport_transform_enabled, 0x64B); -ASSERT_REG_POSITION(view_volume_clip_control, 0x64F); -ASSERT_REG_POSITION(topology_override, 0x65C); -ASSERT_REG_POSITION(depth_bounds_enable, 0x66F); -ASSERT_REG_POSITION(logic_op, 0x671); -ASSERT_REG_POSITION(clear_buffers, 0x674); -ASSERT_REG_POSITION(color_mask, 0x680); -ASSERT_REG_POSITION(query, 0x6C0); -ASSERT_REG_POSITION(vertex_array[0], 0x700); -ASSERT_REG_POSITION(independent_blend, 0x780); -ASSERT_REG_POSITION(vertex_array_limit[0], 0x7C0); -ASSERT_REG_POSITION(shader_config[0], 0x800); -ASSERT_REG_POSITION(firmware, 0x8C0); -ASSERT_REG_POSITION(const_buffer, 0x8E0); -ASSERT_REG_POSITION(cb_bind[0], 0x904); -ASSERT_REG_POSITION(tex_cb_index, 0x982); -ASSERT_REG_POSITION(tfb_varying_locs, 0xA00); -ASSERT_REG_POSITION(ssbo_info, 0xD18); -ASSERT_REG_POSITION(tex_info_buffers.address[0], 0xD2A); -ASSERT_REG_POSITION(tex_info_buffers.size[0], 0xD2F); +ASSERT_REG_POSITION(object_id, 0x0000); +ASSERT_REG_POSITION(nop, 0x0100); +ASSERT_REG_POSITION(notify, 0x0104); +ASSERT_REG_POSITION(wait_for_idle, 0x0110); +ASSERT_REG_POSITION(load_mme, 0x0114); +ASSERT_REG_POSITION(shadow_ram_control, 0x0124); +ASSERT_REG_POSITION(peer, 0x0128); +ASSERT_REG_POSITION(global_render, 0x0130); +ASSERT_REG_POSITION(go_idle, 0x013C); +ASSERT_REG_POSITION(trigger, 0x0140); +ASSERT_REG_POSITION(trigger_wfi, 0x0144); +ASSERT_REG_POSITION(instrumentation_method_header, 0x0150); +ASSERT_REG_POSITION(instrumentation_method_data, 0x0154); +ASSERT_REG_POSITION(upload, 0x0180); +ASSERT_REG_POSITION(launch_dma, 0x01B0); +ASSERT_REG_POSITION(inline_data, 0x01B4); +ASSERT_REG_POSITION(i2m, 0x01DC); +ASSERT_REG_POSITION(run_ds_now, 0x0200); +ASSERT_REG_POSITION(opportunistic_early_z, 0x0204); +ASSERT_REG_POSITION(aliased_line_width_enabled, 0x020C); +ASSERT_REG_POSITION(mandated_early_z, 0x0210); +ASSERT_REG_POSITION(gs_dm_fifo, 0x0214); +ASSERT_REG_POSITION(l2_cache_control, 0x0218); +ASSERT_REG_POSITION(invalidate_shader_cache, 0x021C); +ASSERT_REG_POSITION(sync_info, 0x02C8); +ASSERT_REG_POSITION(prim_circular_buffer_throttle, 0x02D0); +ASSERT_REG_POSITION(flush_invalidate_rop_mini_cache, 0x02D4); +ASSERT_REG_POSITION(surface_clip_block_id, 0x02D8); +ASSERT_REG_POSITION(alpha_circular_buffer_size, 0x02DC); +ASSERT_REG_POSITION(decompress_surface, 0x02E0); +ASSERT_REG_POSITION(zcull_rop_bypass, 0x02E4); +ASSERT_REG_POSITION(zcull_subregion, 0x02E8); +ASSERT_REG_POSITION(raster_bounding_box, 0x02EC); +ASSERT_REG_POSITION(peer_semaphore_release, 0x02F0); +ASSERT_REG_POSITION(iterated_blend_optimization, 0x02F4); +ASSERT_REG_POSITION(zcull_subregion_allocation, 0x02F8); +ASSERT_REG_POSITION(zcull_subregion_algorithm, 0x02FC); +ASSERT_REG_POSITION(ps_output_sample_mask_usage, 0x0300); +ASSERT_REG_POSITION(draw_zero_index, 0x0304); +ASSERT_REG_POSITION(l1_configuration, 0x0308); +ASSERT_REG_POSITION(render_enable_control_load_const_buffer, 0x030C); +ASSERT_REG_POSITION(spa_version, 0x0310); +ASSERT_REG_POSITION(ieee_clean_update, 0x0314); +ASSERT_REG_POSITION(snap_grid, 0x0318); +ASSERT_REG_POSITION(tessellation, 0x0320); +ASSERT_REG_POSITION(sub_tiling_perf, 0x0360); +ASSERT_REG_POSITION(zcull_subregion_report, 0x036C); +ASSERT_REG_POSITION(balanced_primitive_workload, 0x0374); +ASSERT_REG_POSITION(max_patches_per_batch, 0x0378); +ASSERT_REG_POSITION(rasterize_enable, 0x037C); +ASSERT_REG_POSITION(transform_feedback, 0x0380); +ASSERT_REG_POSITION(transform_feedback.controls, 0x0700); +ASSERT_REG_POSITION(raster_input, 0x0740); +ASSERT_REG_POSITION(transform_feedback_enabled, 0x0744); +ASSERT_REG_POSITION(primitive_restart_topology_change_enable, 0x0748); +ASSERT_REG_POSITION(alpha_fraction, 0x074C); +ASSERT_REG_POSITION(hybrid_aa_control, 0x0754); +ASSERT_REG_POSITION(shader_local_memory, 0x077C); +ASSERT_REG_POSITION(color_zero_bandwidth_clear, 0x07A4); +ASSERT_REG_POSITION(z_zero_bandwidth_clear, 0x07A8); +ASSERT_REG_POSITION(isbe_save_restore_program_offset, 0x07AC); +ASSERT_REG_POSITION(zcull_region, 0x07C0); +ASSERT_REG_POSITION(zeta_read_only, 0x07F8); +ASSERT_REG_POSITION(rt, 0x0800); +ASSERT_REG_POSITION(viewport_transform, 0x0A00); +ASSERT_REG_POSITION(viewports, 0x0C00); +ASSERT_REG_POSITION(windows, 0x0D00); +ASSERT_REG_POSITION(clip_id_extent, 0x0D40); +ASSERT_REG_POSITION(max_geometry_instances_per_task, 0x0D60); +ASSERT_REG_POSITION(visible_call_limit, 0x0D64); +ASSERT_REG_POSITION(statistics_count, 0x0D68); +ASSERT_REG_POSITION(clear_rect, 0x0D6C); +ASSERT_REG_POSITION(vertex_buffer, 0x0D74); +ASSERT_REG_POSITION(depth_mode, 0x0D7C); +ASSERT_REG_POSITION(clear_color, 0x0D80); +ASSERT_REG_POSITION(clear_depth, 0x0D90); +ASSERT_REG_POSITION(shader_cache_icache_prefetch, 0x0D94); +ASSERT_REG_POSITION(force_transition_to_beta, 0x0D98); +ASSERT_REG_POSITION(reduce_colour_thresholds, 0x0D9C); +ASSERT_REG_POSITION(clear_stencil, 0x0DA0); +ASSERT_REG_POSITION(invalidate_shader_cache_no_wfi, 0x0DA4); +ASSERT_REG_POSITION(zcull_serialization, 0x0DA8); +ASSERT_REG_POSITION(polygon_mode_front, 0x0DAC); +ASSERT_REG_POSITION(polygon_mode_back, 0x0DB0); +ASSERT_REG_POSITION(polygon_smooth, 0x0DB4); +ASSERT_REG_POSITION(zeta_mark_clean_ieee, 0x0DB8); +ASSERT_REG_POSITION(zcull_dir_format, 0x0DBC); +ASSERT_REG_POSITION(polygon_offset_point_enable, 0x0DC0); +ASSERT_REG_POSITION(polygon_offset_line_enable, 0x0DC4); +ASSERT_REG_POSITION(polygon_offset_fill_enable, 0x0DC8); +ASSERT_REG_POSITION(patch_vertices, 0x0DCC); +ASSERT_REG_POSITION(iterated_blend, 0x0DD0); +ASSERT_REG_POSITION(zcull_criteria, 0x0DD8); +ASSERT_REG_POSITION(fragment_barrier, 0x0DE0); +ASSERT_REG_POSITION(sm_timeout, 0x0DE4); +ASSERT_REG_POSITION(primitive_restart_array, 0x0DE8); +ASSERT_REG_POSITION(load_iterated_blend, 0x0DF0); +ASSERT_REG_POSITION(window_offset_x, 0x0DF8); +ASSERT_REG_POSITION(window_offset_y, 0x0DFC); +ASSERT_REG_POSITION(scissor_test, 0x0E00); +ASSERT_REG_POSITION(select_texture_headers, 0x0F10); +ASSERT_REG_POSITION(vpc_perf, 0x0F14); +ASSERT_REG_POSITION(pm_local_trigger, 0x0F18); +ASSERT_REG_POSITION(post_z_pixel_imask, 0x0F1C); +ASSERT_REG_POSITION(const_color_rendering, 0x0F40); +ASSERT_REG_POSITION(stencil_back_ref, 0x0F54); +ASSERT_REG_POSITION(stencil_back_mask, 0x0F58); +ASSERT_REG_POSITION(stencil_back_func_mask, 0x0F5C); +ASSERT_REG_POSITION(vertex_stream_substitute, 0x0F84); +ASSERT_REG_POSITION(line_mode_clip_generated_edge_do_not_draw, 0x0F8C); +ASSERT_REG_POSITION(color_mask_common, 0x0F90); +ASSERT_REG_POSITION(vtg_warp_watermarks, 0x0F98); +ASSERT_REG_POSITION(depth_bounds, 0x0F9C); +ASSERT_REG_POSITION(sample_mask_target, 0x0FA4); +ASSERT_REG_POSITION(color_target_mrt_enable, 0x0FAC); +ASSERT_REG_POSITION(non_multisampled_z, 0x0FB0); +ASSERT_REG_POSITION(tir_mode, 0x0FB4); +ASSERT_REG_POSITION(anti_alias_raster, 0x0FB8); +ASSERT_REG_POSITION(sample_mask_pos, 0x0FBC); +ASSERT_REG_POSITION(surface_clip_id_memory, 0x0FCC); +ASSERT_REG_POSITION(tir_modulation, 0x0FD4); +ASSERT_REG_POSITION(blend_control_allow_float_pixel_kills, 0x0FDC); +ASSERT_REG_POSITION(zeta, 0x0FE0); +ASSERT_REG_POSITION(surface_clip, 0x0FF4); +ASSERT_REG_POSITION(tiled_cache_treat_heavy_as_light, 0x0FFC); +ASSERT_REG_POSITION(l2_cache_vaf, 0x1000); +ASSERT_REG_POSITION(viewport_multicast, 0x1004); +ASSERT_REG_POSITION(tessellation_cut_height, 0x1008); +ASSERT_REG_POSITION(max_gs_instances_per_task, 0x100C); +ASSERT_REG_POSITION(max_gs_output_vertices_per_task, 0x1010); +ASSERT_REG_POSITION(reserved_sw_method0, 0x1014); +ASSERT_REG_POSITION(gs_output_cb_storage_multiplier, 0x1018); +ASSERT_REG_POSITION(beta_cb_storage_constant, 0x101C); +ASSERT_REG_POSITION(ti_output_cb_storage_multiplier, 0x1020); +ASSERT_REG_POSITION(alpha_cb_storage_constraint, 0x1024); +ASSERT_REG_POSITION(reserved_sw_method1, 0x1028); +ASSERT_REG_POSITION(reserved_sw_method2, 0x102C); +ASSERT_REG_POSITION(tir_modulation_coeff, 0x1030); +ASSERT_REG_POSITION(spare_nop, 0x1044); +ASSERT_REG_POSITION(reserved_sw_method3_to_7, 0x10B0); +ASSERT_REG_POSITION(reduce_color_thresholds_unorm8, 0x10CC); +ASSERT_REG_POSITION(reserved_sw_method10_to_13, 0x10D0); +ASSERT_REG_POSITION(reduce_color_thresholds_unorm10, 0x10E0); +ASSERT_REG_POSITION(reduce_color_thresholds_unorm16, 0x10E4); +ASSERT_REG_POSITION(reduce_color_thresholds_fp11, 0x10E8); +ASSERT_REG_POSITION(reduce_color_thresholds_fp16, 0x10EC); +ASSERT_REG_POSITION(reduce_color_thresholds_srgb8, 0x10F0); +ASSERT_REG_POSITION(unbind_all_constant_buffers, 0x10F4); +ASSERT_REG_POSITION(clear_control, 0x10F8); +ASSERT_REG_POSITION(l2_cache_rop_non_interlocked_reads, 0x10FC); +ASSERT_REG_POSITION(reserved_sw_method14, 0x1100); +ASSERT_REG_POSITION(reserved_sw_method15, 0x1104); +ASSERT_REG_POSITION(no_operation_data_high, 0x110C); +ASSERT_REG_POSITION(depth_bias_control, 0x1110); +ASSERT_REG_POSITION(pm_trigger_end, 0x1114); +ASSERT_REG_POSITION(vertex_id_base, 0x1118); +ASSERT_REG_POSITION(stencil_compression_enabled, 0x111C); +ASSERT_REG_POSITION(vertex_output_attribute_skip_masks, 0x1120); +ASSERT_REG_POSITION(tir_control, 0x1130); +ASSERT_REG_POSITION(mutable_method_treat_mutable_as_heavy, 0x1134); +ASSERT_REG_POSITION(post_ps_use_pre_ps_coverage, 0x1138); +ASSERT_REG_POSITION(fill_via_triangle_mode, 0x113C); +ASSERT_REG_POSITION(blend_per_format_snorm8_unorm16_snorm16_enabled, 0x1140); +ASSERT_REG_POSITION(flush_pending_writes_sm_gloal_store, 0x1144); +ASSERT_REG_POSITION(vertex_attrib_format, 0x1160); +ASSERT_REG_POSITION(multisample_sample_locations, 0x11E0); +ASSERT_REG_POSITION(offset_render_target_index_by_viewport_index, 0x11F0); +ASSERT_REG_POSITION(force_heavy_method_sync, 0x11F4); +ASSERT_REG_POSITION(multisample_coverage_to_color, 0x11F8); +ASSERT_REG_POSITION(decompress_zeta_surface, 0x11FC); +ASSERT_REG_POSITION(zeta_sparse, 0x1208); +ASSERT_REG_POSITION(invalidate_sampler_cache, 0x120C); +ASSERT_REG_POSITION(invalidate_texture_header_cache, 0x1210); +ASSERT_REG_POSITION(vertex_array_instance_first, 0x1214); +ASSERT_REG_POSITION(vertex_array_instance_subsequent, 0x1218); +ASSERT_REG_POSITION(rt_control, 0x121C); +ASSERT_REG_POSITION(compression_threshold_samples, 0x1220); +ASSERT_REG_POSITION(ps_interlock_control, 0x1224); +ASSERT_REG_POSITION(zeta_size, 0x1228); +ASSERT_REG_POSITION(sampler_binding, 0x1234); +ASSERT_REG_POSITION(draw_auto_byte_count, 0x123C); +ASSERT_REG_POSITION(post_vtg_shader_attrib_skip_mask, 0x1240); +ASSERT_REG_POSITION(ps_ticket_dispenser_value, 0x1260); +ASSERT_REG_POSITION(circular_buffer_size, 0x1280); +ASSERT_REG_POSITION(vtg_register_watermarks, 0x1284); +ASSERT_REG_POSITION(invalidate_texture_cache_no_wfi, 0x1288); +ASSERT_REG_POSITION(l2_cache_rop_interlocked_reads, 0x1290); +ASSERT_REG_POSITION(primitive_restart_topology_change_index, 0x12A4); +ASSERT_REG_POSITION(zcull_region_enable, 0x12C8); +ASSERT_REG_POSITION(depth_test_enable, 0x12CC); +ASSERT_REG_POSITION(fill_mode, 0x12D0); +ASSERT_REG_POSITION(shade_mode, 0x12D4); +ASSERT_REG_POSITION(l2_cache_rop_non_interlocked_writes, 0x12D8); +ASSERT_REG_POSITION(l2_cache_rop_interlocked_writes, 0x12DC); +ASSERT_REG_POSITION(alpha_to_coverage_dither, 0x12E0); +ASSERT_REG_POSITION(blend_per_target_enabled, 0x12E4); +ASSERT_REG_POSITION(depth_write_enabled, 0x12E8); +ASSERT_REG_POSITION(alpha_test_enabled, 0x12EC); +ASSERT_REG_POSITION(inline_index_4x8_align, 0x1300); +ASSERT_REG_POSITION(inline_index_4x8_index, 0x1304); +ASSERT_REG_POSITION(d3d_cull_mode, 0x1308); +ASSERT_REG_POSITION(depth_test_func, 0x130C); +ASSERT_REG_POSITION(alpha_test_ref, 0x1310); +ASSERT_REG_POSITION(alpha_test_func, 0x1314); +ASSERT_REG_POSITION(draw_auto_stride, 0x1318); +ASSERT_REG_POSITION(blend_color, 0x131C); +ASSERT_REG_POSITION(invalidate_sampler_cache_lines, 0x1330); +ASSERT_REG_POSITION(invalidate_texture_header_cache_lines, 0x1334); +ASSERT_REG_POSITION(invalidate_texture_data_cache_lines, 0x1338); +ASSERT_REG_POSITION(blend, 0x133C); +ASSERT_REG_POSITION(stencil_enable, 0x1380); +ASSERT_REG_POSITION(stencil_front_op, 0x1384); +ASSERT_REG_POSITION(stencil_front_ref, 0x1394); +ASSERT_REG_POSITION(stencil_front_func_mask, 0x1398); +ASSERT_REG_POSITION(stencil_front_mask, 0x139C); +ASSERT_REG_POSITION(draw_auto_start_byte_count, 0x13A4); +ASSERT_REG_POSITION(frag_color_clamp, 0x13A8); +ASSERT_REG_POSITION(window_origin, 0x13AC); +ASSERT_REG_POSITION(line_width_smooth, 0x13B0); +ASSERT_REG_POSITION(line_width_aliased, 0x13B4); +ASSERT_REG_POSITION(line_override_multisample, 0x1418); +ASSERT_REG_POSITION(alpha_hysteresis_rounds, 0x1420); +ASSERT_REG_POSITION(invalidate_sampler_cache_no_wfi, 0x1424); +ASSERT_REG_POSITION(invalidate_texture_header_cache_no_wfi, 0x1428); +ASSERT_REG_POSITION(global_base_vertex_index, 0x1434); +ASSERT_REG_POSITION(global_base_instance_index, 0x1438); +ASSERT_REG_POSITION(ps_warp_watermarks, 0x1450); +ASSERT_REG_POSITION(ps_regster_watermarks, 0x1454); +ASSERT_REG_POSITION(store_zcull, 0x1464); +ASSERT_REG_POSITION(iterated_blend_constants, 0x1480); +ASSERT_REG_POSITION(load_zcull, 0x1500); +ASSERT_REG_POSITION(surface_clip_id_height, 0x1504); +ASSERT_REG_POSITION(surface_clip_id_clear_rect, 0x1508); +ASSERT_REG_POSITION(user_clip_enable, 0x1510); +ASSERT_REG_POSITION(zpass_pixel_count_enable, 0x1514); +ASSERT_REG_POSITION(point_size, 0x1518); +ASSERT_REG_POSITION(zcull_stats_enable, 0x151C); +ASSERT_REG_POSITION(point_sprite_enable, 0x1520); +ASSERT_REG_POSITION(shader_exceptions_enable, 0x1528); +ASSERT_REG_POSITION(clear_report_value, 0x1530); +ASSERT_REG_POSITION(anti_alias_enable, 0x1534); +ASSERT_REG_POSITION(zeta_enable, 0x1538); +ASSERT_REG_POSITION(anti_alias_alpha_control, 0x153C); +ASSERT_REG_POSITION(render_enable, 0x1550); +ASSERT_REG_POSITION(tex_sampler, 0x155C); +ASSERT_REG_POSITION(slope_scale_depth_bias, 0x156C); +ASSERT_REG_POSITION(line_anti_alias_enable, 0x1570); +ASSERT_REG_POSITION(tex_header, 0x1574); +ASSERT_REG_POSITION(active_zcull_region_id, 0x1590); +ASSERT_REG_POSITION(stencil_two_side_enable, 0x1594); +ASSERT_REG_POSITION(stencil_back_op, 0x1598); +ASSERT_REG_POSITION(framebuffer_srgb, 0x15B8); +ASSERT_REG_POSITION(depth_bias, 0x15BC); +ASSERT_REG_POSITION(zcull_region_format, 0x15C8); +ASSERT_REG_POSITION(rt_layer, 0x15CC); +ASSERT_REG_POSITION(anti_alias_samples_mode, 0x15D0); +ASSERT_REG_POSITION(edge_flag, 0x15E4); +ASSERT_REG_POSITION(draw_inline_index, 0x15E8); +ASSERT_REG_POSITION(inline_index_2x16, 0x15EC); +ASSERT_REG_POSITION(vertex_global_base_offset, 0x15F4); +ASSERT_REG_POSITION(zcull_region_pixel_offset, 0x15FC); +ASSERT_REG_POSITION(point_sprite, 0x1604); +ASSERT_REG_POSITION(program_region, 0x1608); +ASSERT_REG_POSITION(default_attributes, 0x1610); +ASSERT_REG_POSITION(draw, 0x1614); +ASSERT_REG_POSITION(vertex_id_copy, 0x161C); +ASSERT_REG_POSITION(add_to_primitive_id, 0x1620); +ASSERT_REG_POSITION(load_to_primitive_id, 0x1624); +ASSERT_REG_POSITION(shader_based_cull, 0x162C); +ASSERT_REG_POSITION(class_version, 0x1638); +ASSERT_REG_POSITION(primitive_restart, 0x1644); +ASSERT_REG_POSITION(output_vertex_id, 0x164C); +ASSERT_REG_POSITION(anti_alias_point_enable, 0x1658); +ASSERT_REG_POSITION(point_center_mode, 0x165C); +ASSERT_REG_POSITION(line_smooth_params, 0x1668); +ASSERT_REG_POSITION(line_stipple_enable, 0x166C); +ASSERT_REG_POSITION(line_smooth_edge_table, 0x1670); +ASSERT_REG_POSITION(line_stipple_params, 0x1680); +ASSERT_REG_POSITION(provoking_vertex, 0x1684); +ASSERT_REG_POSITION(two_sided_light_enabled, 0x1688); +ASSERT_REG_POSITION(polygon_stipple_enabled, 0x168C); +ASSERT_REG_POSITION(shader_control, 0x1690); +ASSERT_REG_POSITION(class_version_check, 0x16A0); +ASSERT_REG_POSITION(sph_version, 0x16A4); +ASSERT_REG_POSITION(sph_version_check, 0x16A8); +ASSERT_REG_POSITION(alpha_to_coverage_override, 0x16B4); +ASSERT_REG_POSITION(polygon_stipple_pattern, 0x1700); +ASSERT_REG_POSITION(aam_version, 0x1790); +ASSERT_REG_POSITION(aam_version_check, 0x1794); +ASSERT_REG_POSITION(zeta_layer_offset, 0x179C); +ASSERT_REG_POSITION(index_buffer, 0x17C8); +ASSERT_REG_POSITION(index_buffer32_first, 0x17E4); +ASSERT_REG_POSITION(index_buffer16_first, 0x17E8); +ASSERT_REG_POSITION(index_buffer8_first, 0x17EC); +ASSERT_REG_POSITION(index_buffer32_subsequent, 0x17F0); +ASSERT_REG_POSITION(index_buffer16_subsequent, 0x17F4); +ASSERT_REG_POSITION(index_buffer8_subsequent, 0x17F8); +ASSERT_REG_POSITION(depth_bias_clamp, 0x187C); +ASSERT_REG_POSITION(vertex_stream_instances, 0x1880); +ASSERT_REG_POSITION(point_size_attribute, 0x1910); +ASSERT_REG_POSITION(gl_cull_test_enabled, 0x1918); +ASSERT_REG_POSITION(gl_front_face, 0x191C); +ASSERT_REG_POSITION(gl_cull_face, 0x1920); +ASSERT_REG_POSITION(viewport_pixel_center, 0x1924); +ASSERT_REG_POSITION(viewport_scale_offset_enbled, 0x192C); +ASSERT_REG_POSITION(viewport_clip_control, 0x193C); +ASSERT_REG_POSITION(user_clip_op, 0x1940); +ASSERT_REG_POSITION(render_enable_override, 0x1944); +ASSERT_REG_POSITION(primitive_topology_control, 0x1948); +ASSERT_REG_POSITION(window_clip_enable, 0x194C); +ASSERT_REG_POSITION(invalidate_zcull, 0x1958); +ASSERT_REG_POSITION(zcull, 0x1968); +ASSERT_REG_POSITION(topology_override, 0x1970); +ASSERT_REG_POSITION(zcull_sync, 0x1978); +ASSERT_REG_POSITION(clip_id_test_enable, 0x197C); +ASSERT_REG_POSITION(surface_clip_id_width, 0x1980); +ASSERT_REG_POSITION(clip_id, 0x1984); +ASSERT_REG_POSITION(depth_bounds_enable, 0x19BC); +ASSERT_REG_POSITION(blend_float_zero_times_anything_is_zero, 0x19C0); +ASSERT_REG_POSITION(logic_op, 0x19C4); +ASSERT_REG_POSITION(z_compression_enable, 0x19CC); +ASSERT_REG_POSITION(clear_surface, 0x19D0); +ASSERT_REG_POSITION(clear_clip_id_surface, 0x19D4); +ASSERT_REG_POSITION(color_compression_enable, 0x19E0); +ASSERT_REG_POSITION(color_mask, 0x1A00); +ASSERT_REG_POSITION(pipe_nop, 0x1A2C); +ASSERT_REG_POSITION(spare, 0x1A30); +ASSERT_REG_POSITION(report_semaphore, 0x1B00); +ASSERT_REG_POSITION(vertex_streams, 0x1C00); +ASSERT_REG_POSITION(blend_per_target, 0x1E00); +ASSERT_REG_POSITION(vertex_stream_limits, 0x1F00); +ASSERT_REG_POSITION(pipelines, 0x2000); +ASSERT_REG_POSITION(falcon, 0x2300); +ASSERT_REG_POSITION(const_buffer, 0x2380); +ASSERT_REG_POSITION(bind_groups, 0x2400); +ASSERT_REG_POSITION(color_clamp_enable, 0x2600); +ASSERT_REG_POSITION(bindless_texture_const_buffer_slot, 0x2608); +ASSERT_REG_POSITION(trap_handler, 0x260C); +ASSERT_REG_POSITION(stream_out_layout, 0x2800); +ASSERT_REG_POSITION(shader_performance, 0x333C); +ASSERT_REG_POSITION(shadow_scratch, 0x3400); #undef ASSERT_REG_POSITION diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h index 0a4a8b14f..d0709dc69 100644 --- a/src/video_core/gpu.h +++ b/src/video_core/gpu.h @@ -24,11 +24,15 @@ namespace Tegra { class DmaPusher; struct CommandList; +// TODO: Implement the commented ones enum class RenderTargetFormat : u32 { NONE = 0x0, R32B32G32A32_FLOAT = 0xC0, R32G32B32A32_SINT = 0xC1, R32G32B32A32_UINT = 0xC2, + // R32G32B32X32_FLOAT = 0xC3, + // R32G32B32X32_SINT = 0xC4, + // R32G32B32X32_UINT = 0xC5, R16G16B16A16_UNORM = 0xC6, R16G16B16A16_SNORM = 0xC7, R16G16B16A16_SINT = 0xC8, @@ -38,8 +42,8 @@ enum class RenderTargetFormat : u32 { R32G32_SINT = 0xCC, R32G32_UINT = 0xCD, R16G16B16X16_FLOAT = 0xCE, - B8G8R8A8_UNORM = 0xCF, - B8G8R8A8_SRGB = 0xD0, + A8R8G8B8_UNORM = 0xCF, + A8R8G8B8_SRGB = 0xD0, A2B10G10R10_UNORM = 0xD1, A2B10G10R10_UINT = 0xD2, A8B8G8R8_UNORM = 0xD5, @@ -52,10 +56,13 @@ enum class RenderTargetFormat : u32 { R16G16_SINT = 0xDC, R16G16_UINT = 0xDD, R16G16_FLOAT = 0xDE, + // A2R10G10B10_UNORM = 0xDF, B10G11R11_FLOAT = 0xE0, R32_SINT = 0xE3, R32_UINT = 0xE4, R32_FLOAT = 0xE5, + // X8R8G8B8_UNORM = 0xE6, + // X8R8G8B8_SRGB = 0xE7, R5G6B5_UNORM = 0xE8, A1R5G5B5_UNORM = 0xE9, R8G8_UNORM = 0xEA, @@ -71,17 +78,42 @@ enum class RenderTargetFormat : u32 { R8_SNORM = 0xF4, R8_SINT = 0xF5, R8_UINT = 0xF6, + + /* + A8_UNORM = 0xF7, + X1R5G5B5_UNORM = 0xF8, + X8B8G8R8_UNORM = 0xF9, + X8B8G8R8_SRGB = 0xFA, + Z1R5G5B5_UNORM = 0xFB, + O1R5G5B5_UNORM = 0xFC, + Z8R8G8B8_UNORM = 0xFD, + O8R8G8B8_UNORM = 0xFE, + R32_UNORM = 0xFF, + A16_UNORM = 0x40, + A16_FLOAT = 0x41, + A32_FLOAT = 0x42, + A8R8_UNORM = 0x43, + R16A16_UNORM = 0x44, + R16A16_FLOAT = 0x45, + R32A32_FLOAT = 0x46, + B8G8R8A8_UNORM = 0x47, + */ }; enum class DepthFormat : u32 { - D32_FLOAT = 0xA, - D16_UNORM = 0x13, - S8_UINT_Z24_UNORM = 0x14, - D24X8_UNORM = 0x15, - D24S8_UNORM = 0x16, + Z32_FLOAT = 0xA, + Z16_UNORM = 0x13, + Z24_UNORM_S8_UINT = 0x14, + X8Z24_UNORM = 0x15, + S8Z24_UNORM = 0x16, S8_UINT = 0x17, - D24C8_UNORM = 0x18, - D32_FLOAT_S8X24_UINT = 0x19, + V8Z24_UNORM = 0x18, + Z32_FLOAT_X24S8_UINT = 0x19, + /* + X8Z24_UNORM_X16V8S8_UINT = 0x1D, + Z32_FLOAT_X16V8X8_UINT = 0x1E, + Z32_FLOAT_X16V8S8_UINT = 0x1F, + */ }; namespace Engines { diff --git a/src/video_core/host1x/syncpoint_manager.cpp b/src/video_core/host1x/syncpoint_manager.cpp index 326e8355a..a44fc83d3 100644 --- a/src/video_core/host1x/syncpoint_manager.cpp +++ b/src/video_core/host1x/syncpoint_manager.cpp @@ -36,7 +36,17 @@ SyncpointManager::ActionHandle SyncpointManager::RegisterAction( void SyncpointManager::DeregisterAction(std::list<RegisteredAction>& action_storage, ActionHandle& handle) { std::unique_lock lk(guard); - action_storage.erase(handle); + + // We want to ensure the iterator still exists prior to erasing it + // Otherwise, if an invalid iterator was passed in then it could lead to UB + // It is important to avoid UB in that case since the deregister isn't called from a locked + // context + for (auto it = action_storage.begin(); it != action_storage.end(); it++) { + if (it == handle) { + action_storage.erase(it); + return; + } + } } void SyncpointManager::DeregisterGuestAction(u32 syncpoint_id, ActionHandle& handle) { diff --git a/src/video_core/macro/macro_hle.cpp b/src/video_core/macro/macro_hle.cpp index cabe8dcbf..8a8adbb42 100644 --- a/src/video_core/macro/macro_hle.cpp +++ b/src/video_core/macro/macro_hle.cpp @@ -21,16 +21,16 @@ void HLE_771BB18C62444DA0(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& maxwell3d.regs.draw.topology.Assign( static_cast<Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology>(parameters[0] & 0x3ffffff)); - maxwell3d.regs.vb_base_instance = parameters[5]; + maxwell3d.regs.global_base_instance_index = parameters[5]; maxwell3d.mme_draw.instance_count = instance_count; - maxwell3d.regs.vb_element_base = parameters[3]; - maxwell3d.regs.index_array.count = parameters[1]; - maxwell3d.regs.index_array.first = parameters[4]; + maxwell3d.regs.global_base_vertex_index = parameters[3]; + maxwell3d.regs.index_buffer.count = parameters[1]; + maxwell3d.regs.index_buffer.first = parameters[4]; if (maxwell3d.ShouldExecute()) { maxwell3d.Rasterizer().Draw(true, true); } - maxwell3d.regs.index_array.count = 0; + maxwell3d.regs.index_buffer.count = 0; maxwell3d.mme_draw.instance_count = 0; maxwell3d.mme_draw.current_mode = Engines::Maxwell3D::MMEDrawMode::Undefined; } @@ -40,7 +40,7 @@ void HLE_0D61FC9FAAC9FCAD(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& maxwell3d.regs.vertex_buffer.first = parameters[3]; maxwell3d.regs.vertex_buffer.count = parameters[1]; - maxwell3d.regs.vb_base_instance = parameters[4]; + maxwell3d.regs.global_base_instance_index = parameters[4]; maxwell3d.regs.draw.topology.Assign( static_cast<Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology>(parameters[0])); maxwell3d.mme_draw.instance_count = count; @@ -57,12 +57,12 @@ void HLE_0217920100488FF7(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& const u32 instance_count = (maxwell3d.GetRegisterValue(0xD1B) & parameters[2]); const u32 element_base = parameters[4]; const u32 base_instance = parameters[5]; - maxwell3d.regs.index_array.first = parameters[3]; - maxwell3d.regs.reg_array[0x446] = element_base; // vertex id base? - maxwell3d.regs.index_array.count = parameters[1]; + maxwell3d.regs.index_buffer.first = parameters[3]; + maxwell3d.regs.vertex_id_base = element_base; + maxwell3d.regs.index_buffer.count = parameters[1]; maxwell3d.dirty.flags[VideoCommon::Dirty::IndexBuffer] = true; - maxwell3d.regs.vb_element_base = element_base; - maxwell3d.regs.vb_base_instance = base_instance; + maxwell3d.regs.global_base_vertex_index = element_base; + maxwell3d.regs.global_base_instance_index = base_instance; maxwell3d.mme_draw.instance_count = instance_count; maxwell3d.CallMethodFromMME(0x8e3, 0x640); maxwell3d.CallMethodFromMME(0x8e4, element_base); @@ -72,10 +72,10 @@ void HLE_0217920100488FF7(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& if (maxwell3d.ShouldExecute()) { maxwell3d.Rasterizer().Draw(true, true); } - maxwell3d.regs.reg_array[0x446] = 0x0; // vertex id base? - maxwell3d.regs.index_array.count = 0; - maxwell3d.regs.vb_element_base = 0x0; - maxwell3d.regs.vb_base_instance = 0x0; + maxwell3d.regs.vertex_id_base = 0x0; + maxwell3d.regs.index_buffer.count = 0; + maxwell3d.regs.global_base_vertex_index = 0x0; + maxwell3d.regs.global_base_instance_index = 0x0; maxwell3d.mme_draw.instance_count = 0; maxwell3d.CallMethodFromMME(0x8e3, 0x640); maxwell3d.CallMethodFromMME(0x8e4, 0x0); @@ -87,10 +87,10 @@ void HLE_0217920100488FF7(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& void HLE_3F5E74B9C9A50164(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& parameters) { SCOPE_EXIT({ // Clean everything. - maxwell3d.regs.reg_array[0x446] = 0x0; // vertex id base? - maxwell3d.regs.index_array.count = 0; - maxwell3d.regs.vb_element_base = 0x0; - maxwell3d.regs.vb_base_instance = 0x0; + maxwell3d.regs.vertex_id_base = 0x0; + maxwell3d.regs.index_buffer.count = 0; + maxwell3d.regs.global_base_vertex_index = 0x0; + maxwell3d.regs.global_base_instance_index = 0x0; maxwell3d.mme_draw.instance_count = 0; maxwell3d.CallMethodFromMME(0x8e3, 0x640); maxwell3d.CallMethodFromMME(0x8e4, 0x0); @@ -122,11 +122,11 @@ void HLE_3F5E74B9C9A50164(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& const u32 first_index = parameters[base + 2]; const u32 base_vertex = parameters[base + 3]; const u32 base_instance = parameters[base + 4]; - maxwell3d.regs.index_array.first = first_index; - maxwell3d.regs.reg_array[0x446] = base_vertex; - maxwell3d.regs.index_array.count = num_vertices; - maxwell3d.regs.vb_element_base = base_vertex; - maxwell3d.regs.vb_base_instance = base_instance; + maxwell3d.regs.index_buffer.first = first_index; + maxwell3d.regs.vertex_id_base = base_vertex; + maxwell3d.regs.index_buffer.count = num_vertices; + maxwell3d.regs.global_base_vertex_index = base_vertex; + maxwell3d.regs.global_base_instance_index = base_instance; maxwell3d.mme_draw.instance_count = instance_count; maxwell3d.CallMethodFromMME(0x8e3, 0x640); maxwell3d.CallMethodFromMME(0x8e4, base_vertex); diff --git a/src/video_core/query_cache.h b/src/video_core/query_cache.h index b0ebe71b7..00ce53e3e 100644 --- a/src/video_core/query_cache.h +++ b/src/video_core/query_cache.h @@ -137,7 +137,7 @@ public: std::unique_lock lock{mutex}; if (maxwell3d) { const auto& regs = maxwell3d->regs; - Stream(VideoCore::QueryType::SamplesPassed).Update(regs.samplecnt_enable); + Stream(VideoCore::QueryType::SamplesPassed).Update(regs.zpass_pixel_count_enable); } } diff --git a/src/video_core/renderer_base.cpp b/src/video_core/renderer_base.cpp index 45791aa75..e8761a747 100644 --- a/src/video_core/renderer_base.cpp +++ b/src/video_core/renderer_base.cpp @@ -1,6 +1,8 @@ // SPDX-FileCopyrightText: 2015 Citra Emulator Project // SPDX-License-Identifier: GPL-2.0-or-later +#include <thread> + #include "common/logging/log.h" #include "core/frontend/emu_window.h" #include "video_core/renderer_base.h" @@ -35,8 +37,12 @@ void RendererBase::RequestScreenshot(void* data, std::function<void(bool)> callb LOG_ERROR(Render, "A screenshot is already requested or in progress, ignoring the request"); return; } + auto async_callback{[callback = std::move(callback)](bool invert_y) { + std::thread t{callback, invert_y}; + t.detach(); + }}; renderer_settings.screenshot_bits = data; - renderer_settings.screenshot_complete_callback = std::move(callback); + renderer_settings.screenshot_complete_callback = async_callback; renderer_settings.screenshot_framebuffer_layout = layout; renderer_settings.screenshot_requested = true; } diff --git a/src/video_core/renderer_opengl/gl_graphics_pipeline.cpp b/src/video_core/renderer_opengl/gl_graphics_pipeline.cpp index 41493a7da..1d20a79ec 100644 --- a/src/video_core/renderer_opengl/gl_graphics_pipeline.cpp +++ b/src/video_core/renderer_opengl/gl_graphics_pipeline.cpp @@ -73,8 +73,8 @@ GLenum AssemblyStage(size_t stage_index) { /// @param location Hardware location /// @return Pair of ARB_transform_feedback3 token stream first and third arguments /// @note Read https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_transform_feedback3.txt -std::pair<GLint, GLint> TransformFeedbackEnum(u8 location) { - const u8 index = location / 4; +std::pair<GLint, GLint> TransformFeedbackEnum(u32 location) { + const auto index = location / 4; if (index >= 8 && index <= 39) { return {GL_GENERIC_ATTRIB_NV, index - 8}; } @@ -286,7 +286,7 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { buffer_cache.runtime.SetEnableStorageBuffers(use_storage_buffers); const auto& regs{maxwell3d->regs}; - const bool via_header_index{regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex}; + const bool via_header_index{regs.sampler_binding == Maxwell::SamplerBinding::ViaHeaderBinding}; const auto config_stage{[&](size_t stage) LAMBDA_FORCEINLINE { const Shader::Info& info{stage_infos[stage]}; buffer_cache.UnbindGraphicsStorageBuffers(stage); @@ -557,10 +557,25 @@ void GraphicsPipeline::GenerateTransformFeedbackState() { ++current_stream; const auto& locations = key.xfb_state.varyings[feedback]; - std::optional<u8> current_index; + std::optional<u32> current_index; for (u32 offset = 0; offset < layout.varying_count; ++offset) { - const u8 location = locations[offset]; - const u8 index = location / 4; + const auto get_attribute = [&locations](u32 index) -> u32 { + switch (index % 4) { + case 0: + return locations[index / 4].attribute0.Value(); + case 1: + return locations[index / 4].attribute1.Value(); + case 2: + return locations[index / 4].attribute2.Value(); + case 3: + return locations[index / 4].attribute3.Value(); + } + UNREACHABLE(); + return 0; + }; + + const auto attribute{get_attribute(offset)}; + const auto index = attribute / 4U; if (current_index == index) { // Increase number of components of the previous attachment @@ -569,7 +584,7 @@ void GraphicsPipeline::GenerateTransformFeedbackState() { } current_index = index; - std::tie(cursor[0], cursor[2]) = TransformFeedbackEnum(location); + std::tie(cursor[0], cursor[2]) = TransformFeedbackEnum(attribute); cursor[1] = 1; cursor += XFB_ENTRY_STRIDE; } diff --git a/src/video_core/renderer_opengl/gl_graphics_pipeline.h b/src/video_core/renderer_opengl/gl_graphics_pipeline.h index a0f0e63cb..ea53ddb46 100644 --- a/src/video_core/renderer_opengl/gl_graphics_pipeline.h +++ b/src/video_core/renderer_opengl/gl_graphics_pipeline.h @@ -37,8 +37,8 @@ struct GraphicsPipelineKey { BitField<0, 1, u32> xfb_enabled; BitField<1, 1, u32> early_z; BitField<2, 4, Maxwell::PrimitiveTopology> gs_input_topology; - BitField<6, 2, Maxwell::TessellationPrimitive> tessellation_primitive; - BitField<8, 2, Maxwell::TessellationSpacing> tessellation_spacing; + BitField<6, 2, Maxwell::Tessellation::DomainType> tessellation_primitive; + BitField<8, 2, Maxwell::Tessellation::Spacing> tessellation_spacing; BitField<10, 1, u32> tessellation_clockwise; }; std::array<u32, 3> padding; diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index c2d80605d..e5c09a969 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -87,7 +87,7 @@ void RasterizerOpenGL::SyncVertexFormats() { } flags[Dirty::VertexFormat0 + index] = false; - const auto attrib = maxwell3d->regs.vertex_attrib_format[index]; + const auto& attrib = maxwell3d->regs.vertex_attrib_format[index]; const auto gl_index = static_cast<GLuint>(index); // Disable constant attributes. @@ -97,8 +97,8 @@ void RasterizerOpenGL::SyncVertexFormats() { } glEnableVertexAttribArray(gl_index); - if (attrib.type == Maxwell::VertexAttribute::Type::SignedInt || - attrib.type == Maxwell::VertexAttribute::Type::UnsignedInt) { + if (attrib.type == Maxwell::VertexAttribute::Type::SInt || + attrib.type == Maxwell::VertexAttribute::Type::UInt) { glVertexAttribIFormat(gl_index, attrib.ComponentCount(), MaxwellToGL::VertexFormat(attrib), attrib.offset); } else { @@ -125,8 +125,8 @@ void RasterizerOpenGL::SyncVertexInstances() { flags[Dirty::VertexInstance0 + index] = false; const auto gl_index = static_cast<GLuint>(index); - const bool instancing_enabled = regs.instanced_arrays.IsInstancingEnabled(gl_index); - const GLuint divisor = instancing_enabled ? regs.vertex_array[index].divisor : 0; + const bool instancing_enabled = regs.vertex_stream_instances.IsInstancingEnabled(gl_index); + const GLuint divisor = instancing_enabled ? regs.vertex_streams[index].frequency : 0; glVertexBindingDivisor(gl_index, divisor); } } @@ -147,27 +147,27 @@ void RasterizerOpenGL::Clear() { bool use_depth{}; bool use_stencil{}; - if (regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B || - regs.clear_buffers.A) { + if (regs.clear_surface.R || regs.clear_surface.G || regs.clear_surface.B || + regs.clear_surface.A) { use_color = true; - const GLuint index = regs.clear_buffers.RT; + const GLuint index = regs.clear_surface.RT; state_tracker.NotifyColorMask(index); - glColorMaski(index, regs.clear_buffers.R != 0, regs.clear_buffers.G != 0, - regs.clear_buffers.B != 0, regs.clear_buffers.A != 0); + glColorMaski(index, regs.clear_surface.R != 0, regs.clear_surface.G != 0, + regs.clear_surface.B != 0, regs.clear_surface.A != 0); // TODO(Rodrigo): Determine if clamping is used on clears SyncFragmentColorClampState(); SyncFramebufferSRGB(); } - if (regs.clear_buffers.Z) { + if (regs.clear_surface.Z) { ASSERT_MSG(regs.zeta_enable != 0, "Tried to clear Z but buffer is not enabled!"); use_depth = true; state_tracker.NotifyDepthMask(); glDepthMask(GL_TRUE); } - if (regs.clear_buffers.S) { + if (regs.clear_surface.S) { ASSERT_MSG(regs.zeta_enable, "Tried to clear stencil but buffer is not enabled!"); use_stencil = true; } @@ -184,16 +184,16 @@ void RasterizerOpenGL::Clear() { texture_cache.UpdateRenderTargets(true); state_tracker.BindFramebuffer(texture_cache.GetFramebuffer()->Handle()); SyncViewport(); - if (regs.clear_flags.scissor) { + if (regs.clear_control.use_scissor) { SyncScissorTest(); } else { state_tracker.NotifyScissor0(); glDisablei(GL_SCISSOR_TEST, 0); } - UNIMPLEMENTED_IF(regs.clear_flags.viewport); + UNIMPLEMENTED_IF(regs.clear_control.use_viewport_clip0); if (use_color) { - glClearBufferfv(GL_COLOR, regs.clear_buffers.RT, regs.clear_color); + glClearBufferfv(GL_COLOR, regs.clear_surface.RT, regs.clear_color.data()); } if (use_depth && use_stencil) { glClearBufferfi(GL_DEPTH_STENCIL, 0, regs.clear_depth, regs.clear_stencil); @@ -227,14 +227,14 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { const GLenum primitive_mode = MaxwellToGL::PrimitiveTopology(maxwell3d->regs.draw.topology); BeginTransformFeedback(pipeline, primitive_mode); - const GLuint base_instance = static_cast<GLuint>(maxwell3d->regs.vb_base_instance); + const GLuint base_instance = static_cast<GLuint>(maxwell3d->regs.global_base_instance_index); const GLsizei num_instances = static_cast<GLsizei>(is_instanced ? maxwell3d->mme_draw.instance_count : 1); if (is_indexed) { - const GLint base_vertex = static_cast<GLint>(maxwell3d->regs.vb_element_base); - const GLsizei num_vertices = static_cast<GLsizei>(maxwell3d->regs.index_array.count); + const GLint base_vertex = static_cast<GLint>(maxwell3d->regs.global_base_vertex_index); + const GLsizei num_vertices = static_cast<GLsizei>(maxwell3d->regs.index_buffer.count); const GLvoid* const offset = buffer_cache_runtime.IndexOffset(); - const GLenum format = MaxwellToGL::IndexFormat(maxwell3d->regs.index_array.format); + const GLenum format = MaxwellToGL::IndexFormat(maxwell3d->regs.index_buffer.format); if (num_instances == 1 && base_instance == 0 && base_vertex == 0) { glDrawElements(primitive_mode, num_vertices, format, offset); } else if (num_instances == 1 && base_instance == 0) { @@ -555,9 +555,9 @@ void RasterizerOpenGL::SyncViewport() { if (dirty_viewport || dirty_clip_control || flags[Dirty::FrontFace]) { flags[Dirty::FrontFace] = false; - GLenum mode = MaxwellToGL::FrontFace(regs.front_face); + GLenum mode = MaxwellToGL::FrontFace(regs.gl_front_face); bool flip_faces = true; - if (regs.screen_y_control.triangle_rast_flip != 0) { + if (regs.window_origin.flip_y != 0) { flip_faces = !flip_faces; } if (regs.viewport_transform[0].scale_y < 0.0f) { @@ -582,14 +582,15 @@ void RasterizerOpenGL::SyncViewport() { if (regs.viewport_transform[0].scale_y < 0.0f) { flip_y = !flip_y; } - if (regs.screen_y_control.y_negate != 0) { + const bool lower_left{regs.window_origin.mode != Maxwell::WindowOrigin::Mode::UpperLeft}; + if (lower_left) { flip_y = !flip_y; } const bool is_zero_to_one = regs.depth_mode == Maxwell::DepthMode::ZeroToOne; const GLenum origin = flip_y ? GL_UPPER_LEFT : GL_LOWER_LEFT; const GLenum depth = is_zero_to_one ? GL_ZERO_TO_ONE : GL_NEGATIVE_ONE_TO_ONE; state_tracker.ClipControl(origin, depth); - state_tracker.SetYNegate(regs.screen_y_control.y_negate != 0); + state_tracker.SetYNegate(lower_left); } const bool is_rescaling{texture_cache.IsRescaling()}; const float scale = is_rescaling ? Settings::values.resolution_info.up_factor : 1.0f; @@ -657,7 +658,13 @@ void RasterizerOpenGL::SyncDepthClamp() { } flags[Dirty::DepthClampEnabled] = false; - oglEnable(GL_DEPTH_CLAMP, maxwell3d->regs.view_volume_clip_control.depth_clamp_disabled == 0); + bool depth_clamp_disabled{maxwell3d->regs.viewport_clip_control.geometry_clip == + Maxwell::ViewportClipControl::GeometryClip::Passthrough || + maxwell3d->regs.viewport_clip_control.geometry_clip == + Maxwell::ViewportClipControl::GeometryClip::FrustumXYZ || + maxwell3d->regs.viewport_clip_control.geometry_clip == + Maxwell::ViewportClipControl::GeometryClip::FrustumZ}; + oglEnable(GL_DEPTH_CLAMP, !depth_clamp_disabled); } void RasterizerOpenGL::SyncClipEnabled(u32 clip_mask) { @@ -667,7 +674,7 @@ void RasterizerOpenGL::SyncClipEnabled(u32 clip_mask) { } flags[Dirty::ClipDistances] = false; - clip_mask &= maxwell3d->regs.clip_distance_enabled; + clip_mask &= maxwell3d->regs.user_clip_enable.raw; if (clip_mask == last_clip_distance_mask) { return; } @@ -689,9 +696,9 @@ void RasterizerOpenGL::SyncCullMode() { if (flags[Dirty::CullTest]) { flags[Dirty::CullTest] = false; - if (regs.cull_test_enabled) { + if (regs.gl_cull_test_enabled) { glEnable(GL_CULL_FACE); - glCullFace(MaxwellToGL::CullFace(regs.cull_face)); + glCullFace(MaxwellToGL::CullFace(regs.gl_cull_face)); } else { glDisable(GL_CULL_FACE); } @@ -743,19 +750,19 @@ void RasterizerOpenGL::SyncStencilTestState() { const auto& regs = maxwell3d->regs; oglEnable(GL_STENCIL_TEST, regs.stencil_enable); - glStencilFuncSeparate(GL_FRONT, MaxwellToGL::ComparisonOp(regs.stencil_front_func_func), - regs.stencil_front_func_ref, regs.stencil_front_func_mask); - glStencilOpSeparate(GL_FRONT, MaxwellToGL::StencilOp(regs.stencil_front_op_fail), - MaxwellToGL::StencilOp(regs.stencil_front_op_zfail), - MaxwellToGL::StencilOp(regs.stencil_front_op_zpass)); + glStencilFuncSeparate(GL_FRONT, MaxwellToGL::ComparisonOp(regs.stencil_front_op.func), + regs.stencil_front_ref, regs.stencil_front_func_mask); + glStencilOpSeparate(GL_FRONT, MaxwellToGL::StencilOp(regs.stencil_front_op.fail), + MaxwellToGL::StencilOp(regs.stencil_front_op.zfail), + MaxwellToGL::StencilOp(regs.stencil_front_op.zpass)); glStencilMaskSeparate(GL_FRONT, regs.stencil_front_mask); if (regs.stencil_two_side_enable) { - glStencilFuncSeparate(GL_BACK, MaxwellToGL::ComparisonOp(regs.stencil_back_func_func), - regs.stencil_back_func_ref, regs.stencil_back_func_mask); - glStencilOpSeparate(GL_BACK, MaxwellToGL::StencilOp(regs.stencil_back_op_fail), - MaxwellToGL::StencilOp(regs.stencil_back_op_zfail), - MaxwellToGL::StencilOp(regs.stencil_back_op_zpass)); + glStencilFuncSeparate(GL_BACK, MaxwellToGL::ComparisonOp(regs.stencil_back_op.func), + regs.stencil_back_ref, regs.stencil_back_mask); + glStencilOpSeparate(GL_BACK, MaxwellToGL::StencilOp(regs.stencil_back_op.fail), + MaxwellToGL::StencilOp(regs.stencil_back_op.zfail), + MaxwellToGL::StencilOp(regs.stencil_back_op.zpass)); glStencilMaskSeparate(GL_BACK, regs.stencil_back_mask); } else { glStencilFuncSeparate(GL_BACK, GL_ALWAYS, 0, 0xFFFFFFFF); @@ -782,7 +789,7 @@ void RasterizerOpenGL::SyncPolygonModes() { flags[Dirty::PolygonModes] = false; const auto& regs = maxwell3d->regs; - if (regs.fill_rectangle) { + if (regs.fill_via_triangle_mode != Maxwell::FillViaTriangleMode::Disabled) { if (!GLAD_GL_NV_fill_rectangle) { LOG_ERROR(Render_OpenGL, "GL_NV_fill_rectangle used and not supported"); glPolygonMode(GL_FRONT_AND_BACK, GL_FILL); @@ -855,8 +862,8 @@ void RasterizerOpenGL::SyncMultiSampleState() { flags[Dirty::MultisampleControl] = false; const auto& regs = maxwell3d->regs; - oglEnable(GL_SAMPLE_ALPHA_TO_COVERAGE, regs.multisample_control.alpha_to_coverage); - oglEnable(GL_SAMPLE_ALPHA_TO_ONE, regs.multisample_control.alpha_to_one); + oglEnable(GL_SAMPLE_ALPHA_TO_COVERAGE, regs.anti_alias_alpha_control.alpha_to_coverage); + oglEnable(GL_SAMPLE_ALPHA_TO_ONE, regs.anti_alias_alpha_control.alpha_to_one); } void RasterizerOpenGL::SyncFragmentColorClampState() { @@ -866,7 +873,8 @@ void RasterizerOpenGL::SyncFragmentColorClampState() { } flags[Dirty::FragmentClampColor] = false; - glClampColor(GL_CLAMP_FRAGMENT_COLOR, maxwell3d->regs.frag_color_clamp ? GL_TRUE : GL_FALSE); + glClampColor(GL_CLAMP_FRAGMENT_COLOR, + maxwell3d->regs.frag_color_clamp.AnyEnabled() ? GL_TRUE : GL_FALSE); } void RasterizerOpenGL::SyncBlendState() { @@ -886,18 +894,18 @@ void RasterizerOpenGL::SyncBlendState() { } flags[Dirty::BlendStates] = false; - if (!regs.independent_blend_enable) { + if (!regs.blend_per_target_enabled) { if (!regs.blend.enable[0]) { glDisable(GL_BLEND); return; } glEnable(GL_BLEND); - glBlendFuncSeparate(MaxwellToGL::BlendFunc(regs.blend.factor_source_rgb), - MaxwellToGL::BlendFunc(regs.blend.factor_dest_rgb), - MaxwellToGL::BlendFunc(regs.blend.factor_source_a), - MaxwellToGL::BlendFunc(regs.blend.factor_dest_a)); - glBlendEquationSeparate(MaxwellToGL::BlendEquation(regs.blend.equation_rgb), - MaxwellToGL::BlendEquation(regs.blend.equation_a)); + glBlendFuncSeparate(MaxwellToGL::BlendFunc(regs.blend.color_source), + MaxwellToGL::BlendFunc(regs.blend.color_dest), + MaxwellToGL::BlendFunc(regs.blend.alpha_source), + MaxwellToGL::BlendFunc(regs.blend.alpha_dest)); + glBlendEquationSeparate(MaxwellToGL::BlendEquation(regs.blend.color_op), + MaxwellToGL::BlendEquation(regs.blend.alpha_op)); return; } @@ -916,14 +924,13 @@ void RasterizerOpenGL::SyncBlendState() { } glEnablei(GL_BLEND, static_cast<GLuint>(i)); - const auto& src = regs.independent_blend[i]; - glBlendFuncSeparatei(static_cast<GLuint>(i), MaxwellToGL::BlendFunc(src.factor_source_rgb), - MaxwellToGL::BlendFunc(src.factor_dest_rgb), - MaxwellToGL::BlendFunc(src.factor_source_a), - MaxwellToGL::BlendFunc(src.factor_dest_a)); - glBlendEquationSeparatei(static_cast<GLuint>(i), - MaxwellToGL::BlendEquation(src.equation_rgb), - MaxwellToGL::BlendEquation(src.equation_a)); + const auto& src = regs.blend_per_target[i]; + glBlendFuncSeparatei(static_cast<GLuint>(i), MaxwellToGL::BlendFunc(src.color_source), + MaxwellToGL::BlendFunc(src.color_dest), + MaxwellToGL::BlendFunc(src.alpha_source), + MaxwellToGL::BlendFunc(src.alpha_dest)); + glBlendEquationSeparatei(static_cast<GLuint>(i), MaxwellToGL::BlendEquation(src.color_op), + MaxwellToGL::BlendEquation(src.alpha_op)); } } @@ -937,7 +944,7 @@ void RasterizerOpenGL::SyncLogicOpState() { const auto& regs = maxwell3d->regs; if (regs.logic_op.enable) { glEnable(GL_COLOR_LOGIC_OP); - glLogicOp(MaxwellToGL::LogicOp(regs.logic_op.operation)); + glLogicOp(MaxwellToGL::LogicOp(regs.logic_op.op)); } else { glDisable(GL_COLOR_LOGIC_OP); } @@ -996,7 +1003,7 @@ void RasterizerOpenGL::SyncPointState() { flags[Dirty::PointSize] = false; oglEnable(GL_POINT_SPRITE, maxwell3d->regs.point_sprite_enable); - oglEnable(GL_PROGRAM_POINT_SIZE, maxwell3d->regs.vp_point_size.enable); + oglEnable(GL_PROGRAM_POINT_SIZE, maxwell3d->regs.point_size_attribute.enabled); const bool is_rescaling{texture_cache.IsRescaling()}; const float scale = is_rescaling ? Settings::values.resolution_info.up_factor : 1.0f; glPointSize(std::max(1.0f, maxwell3d->regs.point_size * scale)); @@ -1010,8 +1017,8 @@ void RasterizerOpenGL::SyncLineState() { flags[Dirty::LineWidth] = false; const auto& regs = maxwell3d->regs; - oglEnable(GL_LINE_SMOOTH, regs.line_smooth_enable); - glLineWidth(regs.line_smooth_enable ? regs.line_width_smooth : regs.line_width_aliased); + oglEnable(GL_LINE_SMOOTH, regs.line_anti_alias_enable); + glLineWidth(regs.line_anti_alias_enable ? regs.line_width_smooth : regs.line_width_aliased); } void RasterizerOpenGL::SyncPolygonOffset() { @@ -1029,8 +1036,8 @@ void RasterizerOpenGL::SyncPolygonOffset() { if (regs.polygon_offset_fill_enable || regs.polygon_offset_line_enable || regs.polygon_offset_point_enable) { // Hardware divides polygon offset units by two - glPolygonOffsetClamp(regs.polygon_offset_factor, regs.polygon_offset_units / 2.0f, - regs.polygon_offset_clamp); + glPolygonOffsetClamp(regs.slope_scale_depth_bias, regs.depth_bias / 2.0f, + regs.depth_bias_clamp); } } @@ -1062,14 +1069,14 @@ void RasterizerOpenGL::SyncFramebufferSRGB() { void RasterizerOpenGL::BeginTransformFeedback(GraphicsPipeline* program, GLenum primitive_mode) { const auto& regs = maxwell3d->regs; - if (regs.tfb_enabled == 0) { + if (regs.transform_feedback_enabled == 0) { return; } program->ConfigureTransformFeedback(); - UNIMPLEMENTED_IF(regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::TesselationControl) || - regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::TesselationEval) || - regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::Geometry)); + UNIMPLEMENTED_IF(regs.IsShaderConfigEnabled(Maxwell::ShaderType::TessellationInit) || + regs.IsShaderConfigEnabled(Maxwell::ShaderType::Tessellation) || + regs.IsShaderConfigEnabled(Maxwell::ShaderType::Geometry)); UNIMPLEMENTED_IF(primitive_mode != GL_POINTS); // We may have to call BeginTransformFeedbackNV here since they seem to call different @@ -1080,7 +1087,7 @@ void RasterizerOpenGL::BeginTransformFeedback(GraphicsPipeline* program, GLenum } void RasterizerOpenGL::EndTransformFeedback() { - if (maxwell3d->regs.tfb_enabled != 0) { + if (maxwell3d->regs.transform_feedback_enabled != 0) { glEndTransformFeedback(); } } diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 5a29a41d2..6bdb0b645 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -78,11 +78,11 @@ Shader::RuntimeInfo MakeRuntimeInfo(const GraphicsPipelineKey& key, info.tess_clockwise = key.tessellation_clockwise != 0; info.tess_primitive = [&key] { switch (key.tessellation_primitive) { - case Maxwell::TessellationPrimitive::Isolines: + case Maxwell::Tessellation::DomainType::Isolines: return Shader::TessPrimitive::Isolines; - case Maxwell::TessellationPrimitive::Triangles: + case Maxwell::Tessellation::DomainType::Triangles: return Shader::TessPrimitive::Triangles; - case Maxwell::TessellationPrimitive::Quads: + case Maxwell::Tessellation::DomainType::Quads: return Shader::TessPrimitive::Quads; } ASSERT(false); @@ -90,11 +90,11 @@ Shader::RuntimeInfo MakeRuntimeInfo(const GraphicsPipelineKey& key, }(); info.tess_spacing = [&] { switch (key.tessellation_spacing) { - case Maxwell::TessellationSpacing::Equal: + case Maxwell::Tessellation::Spacing::Integer: return Shader::TessSpacing::Equal; - case Maxwell::TessellationSpacing::FractionalOdd: + case Maxwell::Tessellation::Spacing::FractionalOdd: return Shader::TessSpacing::FractionalOdd; - case Maxwell::TessellationSpacing::FractionalEven: + case Maxwell::Tessellation::Spacing::FractionalEven: return Shader::TessSpacing::FractionalEven; } ASSERT(false); @@ -139,14 +139,15 @@ Shader::RuntimeInfo MakeRuntimeInfo(const GraphicsPipelineKey& key, } void SetXfbState(VideoCommon::TransformFeedbackState& state, const Maxwell& regs) { - std::ranges::transform(regs.tfb_layouts, state.layouts.begin(), [](const auto& layout) { - return VideoCommon::TransformFeedbackState::Layout{ - .stream = layout.stream, - .varying_count = layout.varying_count, - .stride = layout.stride, - }; - }); - state.varyings = regs.tfb_varying_locs; + std::ranges::transform(regs.transform_feedback.controls, state.layouts.begin(), + [](const auto& layout) { + return VideoCommon::TransformFeedbackState::Layout{ + .stream = layout.stream, + .varying_count = layout.varying_count, + .stride = layout.stride, + }; + }); + state.varyings = regs.stream_out_layout; } } // Anonymous namespace @@ -309,14 +310,16 @@ GraphicsPipeline* ShaderCache::CurrentGraphicsPipeline() { } const auto& regs{maxwell3d->regs}; graphics_key.raw = 0; - graphics_key.early_z.Assign(regs.force_early_fragment_tests != 0 ? 1 : 0); + graphics_key.early_z.Assign(regs.mandated_early_z != 0 ? 1 : 0); graphics_key.gs_input_topology.Assign(graphics_key.unique_hashes[4] != 0 ? regs.draw.topology.Value() : Maxwell::PrimitiveTopology{}); - graphics_key.tessellation_primitive.Assign(regs.tess_mode.prim.Value()); - graphics_key.tessellation_spacing.Assign(regs.tess_mode.spacing.Value()); - graphics_key.tessellation_clockwise.Assign(regs.tess_mode.cw.Value()); - graphics_key.xfb_enabled.Assign(regs.tfb_enabled != 0 ? 1 : 0); + graphics_key.tessellation_primitive.Assign(regs.tessellation.params.domain_type.Value()); + graphics_key.tessellation_spacing.Assign(regs.tessellation.params.spacing.Value()); + graphics_key.tessellation_clockwise.Assign( + regs.tessellation.params.output_primitives.Value() != + Maxwell::Tessellation::OutputPrimitves::Triangles_CCW); + graphics_key.xfb_enabled.Assign(regs.transform_feedback_enabled != 0 ? 1 : 0); if (graphics_key.xfb_enabled) { SetXfbState(graphics_key.xfb_state, regs); } @@ -354,7 +357,7 @@ GraphicsPipeline* ShaderCache::BuiltPipeline(GraphicsPipeline* pipeline) const n // If games are using a small index count, we can assume these are full screen quads. // Usually these shaders are only used once for building textures so we can assume they // can't be built async - if (maxwell3d->regs.index_array.count <= 6 || maxwell3d->regs.vertex_buffer.count <= 6) { + if (maxwell3d->regs.index_buffer.count <= 6 || maxwell3d->regs.vertex_buffer.count <= 6) { return pipeline; } return nullptr; diff --git a/src/video_core/renderer_opengl/gl_state_tracker.cpp b/src/video_core/renderer_opengl/gl_state_tracker.cpp index a8f3a0f57..a359f96f1 100644 --- a/src/video_core/renderer_opengl/gl_state_tracker.cpp +++ b/src/video_core/renderer_opengl/gl_state_tracker.cpp @@ -38,12 +38,12 @@ void SetupDirtyColorMasks(Tables& tables) { void SetupDirtyVertexInstances(Tables& tables) { static constexpr std::size_t instance_base_offset = 3; for (std::size_t i = 0; i < Regs::NumVertexArrays; ++i) { - const std::size_t array_offset = OFF(vertex_array) + i * NUM(vertex_array[0]); + const std::size_t array_offset = OFF(vertex_streams) + i * NUM(vertex_streams[0]); const std::size_t instance_array_offset = array_offset + instance_base_offset; tables[0][instance_array_offset] = static_cast<u8>(VertexInstance0 + i); tables[1][instance_array_offset] = VertexInstances; - const std::size_t instance_offset = OFF(instanced_arrays) + i; + const std::size_t instance_offset = OFF(vertex_stream_instances) + i; tables[0][instance_offset] = static_cast<u8>(VertexInstance0 + i); tables[1][instance_offset] = VertexInstances; } @@ -70,8 +70,8 @@ void SetupDirtyViewports(Tables& tables) { FillBlock(tables[1], OFF(viewport_transform), NUM(viewport_transform), Viewports); FillBlock(tables[1], OFF(viewports), NUM(viewports), Viewports); - tables[0][OFF(viewport_transform_enabled)] = ViewportTransform; - tables[1][OFF(viewport_transform_enabled)] = Viewports; + tables[0][OFF(viewport_scale_offset_enbled)] = ViewportTransform; + tables[1][OFF(viewport_scale_offset_enbled)] = Viewports; } void SetupDirtyScissors(Tables& tables) { @@ -88,7 +88,7 @@ void SetupDirtyPolygonModes(Tables& tables) { tables[1][OFF(polygon_mode_front)] = PolygonModes; tables[1][OFF(polygon_mode_back)] = PolygonModes; - tables[0][OFF(fill_rectangle)] = PolygonModes; + tables[0][OFF(fill_via_triangle_mode)] = PolygonModes; } void SetupDirtyDepthTest(Tables& tables) { @@ -100,11 +100,11 @@ void SetupDirtyDepthTest(Tables& tables) { void SetupDirtyStencilTest(Tables& tables) { static constexpr std::array offsets = { - OFF(stencil_enable), OFF(stencil_front_func_func), OFF(stencil_front_func_ref), - OFF(stencil_front_func_mask), OFF(stencil_front_op_fail), OFF(stencil_front_op_zfail), - OFF(stencil_front_op_zpass), OFF(stencil_front_mask), OFF(stencil_two_side_enable), - OFF(stencil_back_func_func), OFF(stencil_back_func_ref), OFF(stencil_back_func_mask), - OFF(stencil_back_op_fail), OFF(stencil_back_op_zfail), OFF(stencil_back_op_zpass), + OFF(stencil_enable), OFF(stencil_front_op.func), OFF(stencil_front_ref), + OFF(stencil_front_func_mask), OFF(stencil_front_op.fail), OFF(stencil_front_op.zfail), + OFF(stencil_front_op.zpass), OFF(stencil_front_mask), OFF(stencil_two_side_enable), + OFF(stencil_back_op.func), OFF(stencil_back_ref), OFF(stencil_back_func_mask), + OFF(stencil_back_op.fail), OFF(stencil_back_op.zfail), OFF(stencil_back_op.zpass), OFF(stencil_back_mask)}; for (const auto offset : offsets) { tables[0][offset] = StencilTest; @@ -121,15 +121,15 @@ void SetupDirtyAlphaTest(Tables& tables) { void SetupDirtyBlend(Tables& tables) { FillBlock(tables[0], OFF(blend_color), NUM(blend_color), BlendColor); - tables[0][OFF(independent_blend_enable)] = BlendIndependentEnabled; + tables[0][OFF(blend_per_target_enabled)] = BlendIndependentEnabled; for (std::size_t i = 0; i < Regs::NumRenderTargets; ++i) { - const std::size_t offset = OFF(independent_blend) + i * NUM(independent_blend[0]); - FillBlock(tables[0], offset, NUM(independent_blend[0]), BlendState0 + i); + const std::size_t offset = OFF(blend_per_target) + i * NUM(blend_per_target[0]); + FillBlock(tables[0], offset, NUM(blend_per_target[0]), BlendState0 + i); tables[0][OFF(blend.enable) + i] = static_cast<u8>(BlendState0 + i); } - FillBlock(tables[1], OFF(independent_blend), NUM(independent_blend), BlendStates); + FillBlock(tables[1], OFF(blend_per_target), NUM(blend_per_target), BlendStates); FillBlock(tables[1], OFF(blend), NUM(blend), BlendStates); } @@ -142,13 +142,14 @@ void SetupDirtyPolygonOffset(Tables& tables) { table[OFF(polygon_offset_fill_enable)] = PolygonOffset; table[OFF(polygon_offset_line_enable)] = PolygonOffset; table[OFF(polygon_offset_point_enable)] = PolygonOffset; - table[OFF(polygon_offset_factor)] = PolygonOffset; - table[OFF(polygon_offset_units)] = PolygonOffset; - table[OFF(polygon_offset_clamp)] = PolygonOffset; + table[OFF(slope_scale_depth_bias)] = PolygonOffset; + table[OFF(depth_bias)] = PolygonOffset; + table[OFF(depth_bias_clamp)] = PolygonOffset; } void SetupDirtyMultisampleControl(Tables& tables) { - FillBlock(tables[0], OFF(multisample_control), NUM(multisample_control), MultisampleControl); + FillBlock(tables[0], OFF(anti_alias_alpha_control), NUM(anti_alias_alpha_control), + MultisampleControl); } void SetupDirtyRasterizeEnable(Tables& tables) { @@ -168,7 +169,7 @@ void SetupDirtyFragmentClampColor(Tables& tables) { } void SetupDirtyPointSize(Tables& tables) { - tables[0][OFF(vp_point_size)] = PointSize; + tables[0][OFF(point_size_attribute)] = PointSize; tables[0][OFF(point_size)] = PointSize; tables[0][OFF(point_sprite_enable)] = PointSize; } @@ -176,28 +177,28 @@ void SetupDirtyPointSize(Tables& tables) { void SetupDirtyLineWidth(Tables& tables) { tables[0][OFF(line_width_smooth)] = LineWidth; tables[0][OFF(line_width_aliased)] = LineWidth; - tables[0][OFF(line_smooth_enable)] = LineWidth; + tables[0][OFF(line_anti_alias_enable)] = LineWidth; } void SetupDirtyClipControl(Tables& tables) { auto& table = tables[0]; - table[OFF(screen_y_control)] = ClipControl; + table[OFF(window_origin)] = ClipControl; table[OFF(depth_mode)] = ClipControl; } void SetupDirtyDepthClampEnabled(Tables& tables) { - tables[0][OFF(view_volume_clip_control)] = DepthClampEnabled; + tables[0][OFF(viewport_clip_control)] = DepthClampEnabled; } void SetupDirtyMisc(Tables& tables) { auto& table = tables[0]; - table[OFF(clip_distance_enabled)] = ClipDistances; + table[OFF(user_clip_enable)] = ClipDistances; - table[OFF(front_face)] = FrontFace; + table[OFF(gl_front_face)] = FrontFace; - table[OFF(cull_test_enabled)] = CullTest; - table[OFF(cull_face)] = CullTest; + table[OFF(gl_cull_test_enabled)] = CullTest; + table[OFF(gl_cull_face)] = CullTest; } } // Anonymous namespace diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index 004421236..e14f9b2db 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -126,51 +126,60 @@ inline const FormatTuple& GetFormatTuple(VideoCore::Surface::PixelFormat pixel_f inline GLenum VertexFormat(Maxwell::VertexAttribute attrib) { switch (attrib.type) { - case Maxwell::VertexAttribute::Type::UnsignedNorm: - case Maxwell::VertexAttribute::Type::UnsignedScaled: - case Maxwell::VertexAttribute::Type::UnsignedInt: + case Maxwell::VertexAttribute::Type::UnusedEnumDoNotUseBecauseItWillGoAway: + ASSERT_MSG(false, "Invalid vertex attribute type!"); + break; + case Maxwell::VertexAttribute::Type::UNorm: + case Maxwell::VertexAttribute::Type::UScaled: + case Maxwell::VertexAttribute::Type::UInt: switch (attrib.size) { - case Maxwell::VertexAttribute::Size::Size_8: - case Maxwell::VertexAttribute::Size::Size_8_8: - case Maxwell::VertexAttribute::Size::Size_8_8_8: - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8: + case Maxwell::VertexAttribute::Size::Size_A8: + case Maxwell::VertexAttribute::Size::Size_R8_G8: + case Maxwell::VertexAttribute::Size::Size_G8_R8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8_A8: + case Maxwell::VertexAttribute::Size::Size_X8_B8_G8_R8: return GL_UNSIGNED_BYTE; - case Maxwell::VertexAttribute::Size::Size_16: - case Maxwell::VertexAttribute::Size::Size_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return GL_UNSIGNED_SHORT; - case Maxwell::VertexAttribute::Size::Size_32: - case Maxwell::VertexAttribute::Size::Size_32_32: - case Maxwell::VertexAttribute::Size::Size_32_32_32: - case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32: + case Maxwell::VertexAttribute::Size::Size_R32_G32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32_A32: return GL_UNSIGNED_INT; - case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + case Maxwell::VertexAttribute::Size::Size_A2_B10_G10_R10: return GL_UNSIGNED_INT_2_10_10_10_REV; default: break; } break; - case Maxwell::VertexAttribute::Type::SignedNorm: - case Maxwell::VertexAttribute::Type::SignedScaled: - case Maxwell::VertexAttribute::Type::SignedInt: + case Maxwell::VertexAttribute::Type::SNorm: + case Maxwell::VertexAttribute::Type::SScaled: + case Maxwell::VertexAttribute::Type::SInt: switch (attrib.size) { - case Maxwell::VertexAttribute::Size::Size_8: - case Maxwell::VertexAttribute::Size::Size_8_8: - case Maxwell::VertexAttribute::Size::Size_8_8_8: - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8: + case Maxwell::VertexAttribute::Size::Size_A8: + case Maxwell::VertexAttribute::Size::Size_R8_G8: + case Maxwell::VertexAttribute::Size::Size_G8_R8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8_A8: + case Maxwell::VertexAttribute::Size::Size_X8_B8_G8_R8: return GL_BYTE; - case Maxwell::VertexAttribute::Size::Size_16: - case Maxwell::VertexAttribute::Size::Size_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return GL_SHORT; - case Maxwell::VertexAttribute::Size::Size_32: - case Maxwell::VertexAttribute::Size::Size_32_32: - case Maxwell::VertexAttribute::Size::Size_32_32_32: - case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32: + case Maxwell::VertexAttribute::Size::Size_R32_G32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32_A32: return GL_INT; - case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + case Maxwell::VertexAttribute::Size::Size_A2_B10_G10_R10: return GL_INT_2_10_10_10_REV; default: break; @@ -178,17 +187,17 @@ inline GLenum VertexFormat(Maxwell::VertexAttribute attrib) { break; case Maxwell::VertexAttribute::Type::Float: switch (attrib.size) { - case Maxwell::VertexAttribute::Size::Size_16: - case Maxwell::VertexAttribute::Size::Size_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16: - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return GL_HALF_FLOAT; - case Maxwell::VertexAttribute::Size::Size_32: - case Maxwell::VertexAttribute::Size::Size_32_32: - case Maxwell::VertexAttribute::Size::Size_32_32_32: - case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32: + case Maxwell::VertexAttribute::Size::Size_R32_G32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32_A32: return GL_FLOAT; - case Maxwell::VertexAttribute::Size::Size_11_11_10: + case Maxwell::VertexAttribute::Size::Size_B10_G11_R11: return GL_UNSIGNED_INT_10F_11F_11F_REV; default: break; @@ -335,20 +344,20 @@ inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) { inline GLenum BlendEquation(Maxwell::Blend::Equation equation) { switch (equation) { - case Maxwell::Blend::Equation::Add: - case Maxwell::Blend::Equation::AddGL: + case Maxwell::Blend::Equation::Add_D3D: + case Maxwell::Blend::Equation::Add_GL: return GL_FUNC_ADD; - case Maxwell::Blend::Equation::Subtract: - case Maxwell::Blend::Equation::SubtractGL: + case Maxwell::Blend::Equation::Subtract_D3D: + case Maxwell::Blend::Equation::Subtract_GL: return GL_FUNC_SUBTRACT; - case Maxwell::Blend::Equation::ReverseSubtract: - case Maxwell::Blend::Equation::ReverseSubtractGL: + case Maxwell::Blend::Equation::ReverseSubtract_D3D: + case Maxwell::Blend::Equation::ReverseSubtract_GL: return GL_FUNC_REVERSE_SUBTRACT; - case Maxwell::Blend::Equation::Min: - case Maxwell::Blend::Equation::MinGL: + case Maxwell::Blend::Equation::Min_D3D: + case Maxwell::Blend::Equation::Min_GL: return GL_MIN; - case Maxwell::Blend::Equation::Max: - case Maxwell::Blend::Equation::MaxGL: + case Maxwell::Blend::Equation::Max_D3D: + case Maxwell::Blend::Equation::Max_GL: return GL_MAX; } UNIMPLEMENTED_MSG("Unimplemented blend equation={}", equation); @@ -357,62 +366,62 @@ inline GLenum BlendEquation(Maxwell::Blend::Equation equation) { inline GLenum BlendFunc(Maxwell::Blend::Factor factor) { switch (factor) { - case Maxwell::Blend::Factor::Zero: - case Maxwell::Blend::Factor::ZeroGL: + case Maxwell::Blend::Factor::Zero_D3D: + case Maxwell::Blend::Factor::Zero_GL: return GL_ZERO; - case Maxwell::Blend::Factor::One: - case Maxwell::Blend::Factor::OneGL: + case Maxwell::Blend::Factor::One_D3D: + case Maxwell::Blend::Factor::One_GL: return GL_ONE; - case Maxwell::Blend::Factor::SourceColor: - case Maxwell::Blend::Factor::SourceColorGL: + case Maxwell::Blend::Factor::SourceColor_D3D: + case Maxwell::Blend::Factor::SourceColor_GL: return GL_SRC_COLOR; - case Maxwell::Blend::Factor::OneMinusSourceColor: - case Maxwell::Blend::Factor::OneMinusSourceColorGL: + case Maxwell::Blend::Factor::OneMinusSourceColor_D3D: + case Maxwell::Blend::Factor::OneMinusSourceColor_GL: return GL_ONE_MINUS_SRC_COLOR; - case Maxwell::Blend::Factor::SourceAlpha: - case Maxwell::Blend::Factor::SourceAlphaGL: + case Maxwell::Blend::Factor::SourceAlpha_D3D: + case Maxwell::Blend::Factor::SourceAlpha_GL: return GL_SRC_ALPHA; - case Maxwell::Blend::Factor::OneMinusSourceAlpha: - case Maxwell::Blend::Factor::OneMinusSourceAlphaGL: + case Maxwell::Blend::Factor::OneMinusSourceAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusSourceAlpha_GL: return GL_ONE_MINUS_SRC_ALPHA; - case Maxwell::Blend::Factor::DestAlpha: - case Maxwell::Blend::Factor::DestAlphaGL: + case Maxwell::Blend::Factor::DestAlpha_D3D: + case Maxwell::Blend::Factor::DestAlpha_GL: return GL_DST_ALPHA; - case Maxwell::Blend::Factor::OneMinusDestAlpha: - case Maxwell::Blend::Factor::OneMinusDestAlphaGL: + case Maxwell::Blend::Factor::OneMinusDestAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusDestAlpha_GL: return GL_ONE_MINUS_DST_ALPHA; - case Maxwell::Blend::Factor::DestColor: - case Maxwell::Blend::Factor::DestColorGL: + case Maxwell::Blend::Factor::DestColor_D3D: + case Maxwell::Blend::Factor::DestColor_GL: return GL_DST_COLOR; - case Maxwell::Blend::Factor::OneMinusDestColor: - case Maxwell::Blend::Factor::OneMinusDestColorGL: + case Maxwell::Blend::Factor::OneMinusDestColor_D3D: + case Maxwell::Blend::Factor::OneMinusDestColor_GL: return GL_ONE_MINUS_DST_COLOR; - case Maxwell::Blend::Factor::SourceAlphaSaturate: - case Maxwell::Blend::Factor::SourceAlphaSaturateGL: + case Maxwell::Blend::Factor::SourceAlphaSaturate_D3D: + case Maxwell::Blend::Factor::SourceAlphaSaturate_GL: return GL_SRC_ALPHA_SATURATE; - case Maxwell::Blend::Factor::Source1Color: - case Maxwell::Blend::Factor::Source1ColorGL: + case Maxwell::Blend::Factor::Source1Color_D3D: + case Maxwell::Blend::Factor::Source1Color_GL: return GL_SRC1_COLOR; - case Maxwell::Blend::Factor::OneMinusSource1Color: - case Maxwell::Blend::Factor::OneMinusSource1ColorGL: + case Maxwell::Blend::Factor::OneMinusSource1Color_D3D: + case Maxwell::Blend::Factor::OneMinusSource1Color_GL: return GL_ONE_MINUS_SRC1_COLOR; - case Maxwell::Blend::Factor::Source1Alpha: - case Maxwell::Blend::Factor::Source1AlphaGL: + case Maxwell::Blend::Factor::Source1Alpha_D3D: + case Maxwell::Blend::Factor::Source1Alpha_GL: return GL_SRC1_ALPHA; - case Maxwell::Blend::Factor::OneMinusSource1Alpha: - case Maxwell::Blend::Factor::OneMinusSource1AlphaGL: + case Maxwell::Blend::Factor::OneMinusSource1Alpha_D3D: + case Maxwell::Blend::Factor::OneMinusSource1Alpha_GL: return GL_ONE_MINUS_SRC1_ALPHA; - case Maxwell::Blend::Factor::ConstantColor: - case Maxwell::Blend::Factor::ConstantColorGL: + case Maxwell::Blend::Factor::BlendFactor_D3D: + case Maxwell::Blend::Factor::ConstantColor_GL: return GL_CONSTANT_COLOR; - case Maxwell::Blend::Factor::OneMinusConstantColor: - case Maxwell::Blend::Factor::OneMinusConstantColorGL: + case Maxwell::Blend::Factor::OneMinusBlendFactor_D3D: + case Maxwell::Blend::Factor::OneMinusConstantColor_GL: return GL_ONE_MINUS_CONSTANT_COLOR; - case Maxwell::Blend::Factor::ConstantAlpha: - case Maxwell::Blend::Factor::ConstantAlphaGL: + case Maxwell::Blend::Factor::BothSourceAlpha_D3D: + case Maxwell::Blend::Factor::ConstantAlpha_GL: return GL_CONSTANT_ALPHA; - case Maxwell::Blend::Factor::OneMinusConstantAlpha: - case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: + case Maxwell::Blend::Factor::OneMinusBothSourceAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusConstantAlpha_GL: return GL_ONE_MINUS_CONSTANT_ALPHA; } UNIMPLEMENTED_MSG("Unimplemented blend factor={}", factor); @@ -421,60 +430,60 @@ inline GLenum BlendFunc(Maxwell::Blend::Factor factor) { inline GLenum ComparisonOp(Maxwell::ComparisonOp comparison) { switch (comparison) { - case Maxwell::ComparisonOp::Never: - case Maxwell::ComparisonOp::NeverOld: + case Maxwell::ComparisonOp::Never_D3D: + case Maxwell::ComparisonOp::Never_GL: return GL_NEVER; - case Maxwell::ComparisonOp::Less: - case Maxwell::ComparisonOp::LessOld: + case Maxwell::ComparisonOp::Less_D3D: + case Maxwell::ComparisonOp::Less_GL: return GL_LESS; - case Maxwell::ComparisonOp::Equal: - case Maxwell::ComparisonOp::EqualOld: + case Maxwell::ComparisonOp::Equal_D3D: + case Maxwell::ComparisonOp::Equal_GL: return GL_EQUAL; - case Maxwell::ComparisonOp::LessEqual: - case Maxwell::ComparisonOp::LessEqualOld: + case Maxwell::ComparisonOp::LessEqual_D3D: + case Maxwell::ComparisonOp::LessEqual_GL: return GL_LEQUAL; - case Maxwell::ComparisonOp::Greater: - case Maxwell::ComparisonOp::GreaterOld: + case Maxwell::ComparisonOp::Greater_D3D: + case Maxwell::ComparisonOp::Greater_GL: return GL_GREATER; - case Maxwell::ComparisonOp::NotEqual: - case Maxwell::ComparisonOp::NotEqualOld: + case Maxwell::ComparisonOp::NotEqual_D3D: + case Maxwell::ComparisonOp::NotEqual_GL: return GL_NOTEQUAL; - case Maxwell::ComparisonOp::GreaterEqual: - case Maxwell::ComparisonOp::GreaterEqualOld: + case Maxwell::ComparisonOp::GreaterEqual_D3D: + case Maxwell::ComparisonOp::GreaterEqual_GL: return GL_GEQUAL; - case Maxwell::ComparisonOp::Always: - case Maxwell::ComparisonOp::AlwaysOld: + case Maxwell::ComparisonOp::Always_D3D: + case Maxwell::ComparisonOp::Always_GL: return GL_ALWAYS; } UNIMPLEMENTED_MSG("Unimplemented comparison op={}", comparison); return GL_ALWAYS; } -inline GLenum StencilOp(Maxwell::StencilOp stencil) { +inline GLenum StencilOp(Maxwell::StencilOp::Op stencil) { switch (stencil) { - case Maxwell::StencilOp::Keep: - case Maxwell::StencilOp::KeepOGL: + case Maxwell::StencilOp::Op::Keep_D3D: + case Maxwell::StencilOp::Op::Keep_GL: return GL_KEEP; - case Maxwell::StencilOp::Zero: - case Maxwell::StencilOp::ZeroOGL: + case Maxwell::StencilOp::Op::Zero_D3D: + case Maxwell::StencilOp::Op::Zero_GL: return GL_ZERO; - case Maxwell::StencilOp::Replace: - case Maxwell::StencilOp::ReplaceOGL: + case Maxwell::StencilOp::Op::Replace_D3D: + case Maxwell::StencilOp::Op::Replace_GL: return GL_REPLACE; - case Maxwell::StencilOp::Incr: - case Maxwell::StencilOp::IncrOGL: + case Maxwell::StencilOp::Op::IncrSaturate_D3D: + case Maxwell::StencilOp::Op::IncrSaturate_GL: return GL_INCR; - case Maxwell::StencilOp::Decr: - case Maxwell::StencilOp::DecrOGL: + case Maxwell::StencilOp::Op::DecrSaturate_D3D: + case Maxwell::StencilOp::Op::DecrSaturate_GL: return GL_DECR; - case Maxwell::StencilOp::Invert: - case Maxwell::StencilOp::InvertOGL: + case Maxwell::StencilOp::Op::Invert_D3D: + case Maxwell::StencilOp::Op::Invert_GL: return GL_INVERT; - case Maxwell::StencilOp::IncrWrap: - case Maxwell::StencilOp::IncrWrapOGL: + case Maxwell::StencilOp::Op::Incr_D3D: + case Maxwell::StencilOp::Op::Incr_GL: return GL_INCR_WRAP; - case Maxwell::StencilOp::DecrWrap: - case Maxwell::StencilOp::DecrWrapOGL: + case Maxwell::StencilOp::Op::Decr_D3D: + case Maxwell::StencilOp::Op::Decr_GL: return GL_DECR_WRAP; } UNIMPLEMENTED_MSG("Unimplemented stencil op={}", stencil); @@ -505,39 +514,39 @@ inline GLenum CullFace(Maxwell::CullFace cull_face) { return GL_BACK; } -inline GLenum LogicOp(Maxwell::LogicOperation operation) { +inline GLenum LogicOp(Maxwell::LogicOp::Op operation) { switch (operation) { - case Maxwell::LogicOperation::Clear: + case Maxwell::LogicOp::Op::Clear: return GL_CLEAR; - case Maxwell::LogicOperation::And: + case Maxwell::LogicOp::Op::And: return GL_AND; - case Maxwell::LogicOperation::AndReverse: + case Maxwell::LogicOp::Op::AndReverse: return GL_AND_REVERSE; - case Maxwell::LogicOperation::Copy: + case Maxwell::LogicOp::Op::Copy: return GL_COPY; - case Maxwell::LogicOperation::AndInverted: + case Maxwell::LogicOp::Op::AndInverted: return GL_AND_INVERTED; - case Maxwell::LogicOperation::NoOp: + case Maxwell::LogicOp::Op::NoOp: return GL_NOOP; - case Maxwell::LogicOperation::Xor: + case Maxwell::LogicOp::Op::Xor: return GL_XOR; - case Maxwell::LogicOperation::Or: + case Maxwell::LogicOp::Op::Or: return GL_OR; - case Maxwell::LogicOperation::Nor: + case Maxwell::LogicOp::Op::Nor: return GL_NOR; - case Maxwell::LogicOperation::Equiv: + case Maxwell::LogicOp::Op::Equiv: return GL_EQUIV; - case Maxwell::LogicOperation::Invert: + case Maxwell::LogicOp::Op::Invert: return GL_INVERT; - case Maxwell::LogicOperation::OrReverse: + case Maxwell::LogicOp::Op::OrReverse: return GL_OR_REVERSE; - case Maxwell::LogicOperation::CopyInverted: + case Maxwell::LogicOp::Op::CopyInverted: return GL_COPY_INVERTED; - case Maxwell::LogicOperation::OrInverted: + case Maxwell::LogicOp::Op::OrInverted: return GL_OR_INVERTED; - case Maxwell::LogicOperation::Nand: + case Maxwell::LogicOp::Op::Nand: return GL_NAND; - case Maxwell::LogicOperation::Set: + case Maxwell::LogicOp::Op::Set: return GL_SET; } UNIMPLEMENTED_MSG("Unimplemented logic operation={}", operation); diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp index 733b454de..eab1b8f93 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp @@ -34,14 +34,15 @@ constexpr std::array POLYGON_OFFSET_ENABLE_LUT = { }; void RefreshXfbState(VideoCommon::TransformFeedbackState& state, const Maxwell& regs) { - std::ranges::transform(regs.tfb_layouts, state.layouts.begin(), [](const auto& layout) { - return VideoCommon::TransformFeedbackState::Layout{ - .stream = layout.stream, - .varying_count = layout.varying_count, - .stride = layout.stride, - }; - }); - state.varyings = regs.tfb_varying_locs; + std::ranges::transform(regs.transform_feedback.controls, state.layouts.begin(), + [](const auto& layout) { + return VideoCommon::TransformFeedbackState::Layout{ + .stream = layout.stream, + .varying_count = layout.varying_count, + .stride = layout.stride, + }; + }); + state.varyings = regs.stream_out_layout; } } // Anonymous namespace @@ -58,33 +59,38 @@ void FixedPipelineState::Refresh(Tegra::Engines::Maxwell3D& maxwell3d, raw1 = 0; extended_dynamic_state.Assign(has_extended_dynamic_state ? 1 : 0); dynamic_vertex_input.Assign(has_dynamic_vertex_input ? 1 : 0); - xfb_enabled.Assign(regs.tfb_enabled != 0); + xfb_enabled.Assign(regs.transform_feedback_enabled != 0); primitive_restart_enable.Assign(regs.primitive_restart.enabled != 0 ? 1 : 0); depth_bias_enable.Assign(enabled_lut[POLYGON_OFFSET_ENABLE_LUT[topology_index]] != 0 ? 1 : 0); - depth_clamp_disabled.Assign(regs.view_volume_clip_control.depth_clamp_disabled.Value()); + depth_clamp_disabled.Assign(regs.viewport_clip_control.geometry_clip == + Maxwell::ViewportClipControl::GeometryClip::Passthrough || + regs.viewport_clip_control.geometry_clip == + Maxwell::ViewportClipControl::GeometryClip::FrustumXYZ || + regs.viewport_clip_control.geometry_clip == + Maxwell::ViewportClipControl::GeometryClip::FrustumZ); ndc_minus_one_to_one.Assign(regs.depth_mode == Maxwell::DepthMode::MinusOneToOne ? 1 : 0); polygon_mode.Assign(PackPolygonMode(regs.polygon_mode_front)); patch_control_points_minus_one.Assign(regs.patch_vertices - 1); - tessellation_primitive.Assign(static_cast<u32>(regs.tess_mode.prim.Value())); - tessellation_spacing.Assign(static_cast<u32>(regs.tess_mode.spacing.Value())); - tessellation_clockwise.Assign(regs.tess_mode.cw.Value()); + tessellation_primitive.Assign(static_cast<u32>(regs.tessellation.params.domain_type.Value())); + tessellation_spacing.Assign(static_cast<u32>(regs.tessellation.params.spacing.Value())); + tessellation_clockwise.Assign(regs.tessellation.params.output_primitives.Value() != + Maxwell::Tessellation::OutputPrimitves::Triangles_CCW); logic_op_enable.Assign(regs.logic_op.enable != 0 ? 1 : 0); - logic_op.Assign(PackLogicOp(regs.logic_op.operation)); + logic_op.Assign(PackLogicOp(regs.logic_op.op)); topology.Assign(regs.draw.topology); - msaa_mode.Assign(regs.multisample_mode); + msaa_mode.Assign(regs.anti_alias_samples_mode); raw2 = 0; rasterize_enable.Assign(regs.rasterize_enable != 0 ? 1 : 0); const auto test_func = - regs.alpha_test_enabled != 0 ? regs.alpha_test_func : Maxwell::ComparisonOp::Always; + regs.alpha_test_enabled != 0 ? regs.alpha_test_func : Maxwell::ComparisonOp::Always_GL; alpha_test_func.Assign(PackComparisonOp(test_func)); - early_z.Assign(regs.force_early_fragment_tests != 0 ? 1 : 0); + early_z.Assign(regs.mandated_early_z != 0 ? 1 : 0); depth_enabled.Assign(regs.zeta_enable != 0 ? 1 : 0); depth_format.Assign(static_cast<u32>(regs.zeta.format)); - y_negate.Assign(regs.screen_y_control.y_negate != 0 ? 1 : 0); - provoking_vertex_last.Assign(regs.provoking_vertex_last != 0 ? 1 : 0); - conservative_raster_enable.Assign(regs.conservative_raster_enable != 0 ? 1 : 0); - smooth_lines.Assign(regs.line_smooth_enable != 0 ? 1 : 0); + y_negate.Assign(regs.window_origin.mode != Maxwell::WindowOrigin::Mode::UpperLeft ? 1 : 0); + provoking_vertex_last.Assign(regs.provoking_vertex == Maxwell::ProvokingVertex::Last ? 1 : 0); + smooth_lines.Assign(regs.line_anti_alias_enable != 0 ? 1 : 0); for (size_t i = 0; i < regs.rt.size(); ++i) { color_formats[i] = static_cast<u8>(regs.rt[i].format); @@ -116,8 +122,8 @@ void FixedPipelineState::Refresh(Tegra::Engines::Maxwell3D& maxwell3d, maxwell3d.dirty.flags[Dirty::VertexInput] = false; enabled_divisors = 0; for (size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { - const bool is_enabled = regs.instanced_arrays.IsInstancingEnabled(index); - binding_divisors[index] = is_enabled ? regs.vertex_array[index].divisor : 0; + const bool is_enabled = regs.vertex_stream_instances.IsInstancingEnabled(index); + binding_divisors[index] = is_enabled ? regs.vertex_streams[index].frequency : 0; enabled_divisors |= (is_enabled ? u64{1} : 0) << index; } for (size_t index = 0; index < Maxwell::NumVertexAttributes; ++index) { @@ -164,17 +170,17 @@ void FixedPipelineState::BlendingAttachment::Refresh(const Maxwell& regs, size_t // TODO: C++20 Use templated lambda to deduplicate code - if (!regs.independent_blend_enable) { - const auto& src = regs.blend; - if (!src.enable[index]) { + if (!regs.blend_per_target_enabled) { + if (!regs.blend.enable[index]) { return; } - equation_rgb.Assign(PackBlendEquation(src.equation_rgb)); - equation_a.Assign(PackBlendEquation(src.equation_a)); - factor_source_rgb.Assign(PackBlendFactor(src.factor_source_rgb)); - factor_dest_rgb.Assign(PackBlendFactor(src.factor_dest_rgb)); - factor_source_a.Assign(PackBlendFactor(src.factor_source_a)); - factor_dest_a.Assign(PackBlendFactor(src.factor_dest_a)); + const auto& src = regs.blend; + equation_rgb.Assign(PackBlendEquation(src.color_op)); + equation_a.Assign(PackBlendEquation(src.alpha_op)); + factor_source_rgb.Assign(PackBlendFactor(src.color_source)); + factor_dest_rgb.Assign(PackBlendFactor(src.color_dest)); + factor_source_a.Assign(PackBlendFactor(src.alpha_source)); + factor_dest_a.Assign(PackBlendFactor(src.alpha_dest)); enable.Assign(1); return; } @@ -182,34 +188,34 @@ void FixedPipelineState::BlendingAttachment::Refresh(const Maxwell& regs, size_t if (!regs.blend.enable[index]) { return; } - const auto& src = regs.independent_blend[index]; - equation_rgb.Assign(PackBlendEquation(src.equation_rgb)); - equation_a.Assign(PackBlendEquation(src.equation_a)); - factor_source_rgb.Assign(PackBlendFactor(src.factor_source_rgb)); - factor_dest_rgb.Assign(PackBlendFactor(src.factor_dest_rgb)); - factor_source_a.Assign(PackBlendFactor(src.factor_source_a)); - factor_dest_a.Assign(PackBlendFactor(src.factor_dest_a)); + const auto& src = regs.blend_per_target[index]; + equation_rgb.Assign(PackBlendEquation(src.color_op)); + equation_a.Assign(PackBlendEquation(src.alpha_op)); + factor_source_rgb.Assign(PackBlendFactor(src.color_source)); + factor_dest_rgb.Assign(PackBlendFactor(src.color_dest)); + factor_source_a.Assign(PackBlendFactor(src.alpha_source)); + factor_dest_a.Assign(PackBlendFactor(src.alpha_dest)); enable.Assign(1); } void FixedPipelineState::DynamicState::Refresh(const Maxwell& regs) { - u32 packed_front_face = PackFrontFace(regs.front_face); - if (regs.screen_y_control.triangle_rast_flip != 0) { + u32 packed_front_face = PackFrontFace(regs.gl_front_face); + if (regs.window_origin.flip_y != 0) { // Flip front face packed_front_face = 1 - packed_front_face; } raw1 = 0; raw2 = 0; - front.action_stencil_fail.Assign(PackStencilOp(regs.stencil_front_op_fail)); - front.action_depth_fail.Assign(PackStencilOp(regs.stencil_front_op_zfail)); - front.action_depth_pass.Assign(PackStencilOp(regs.stencil_front_op_zpass)); - front.test_func.Assign(PackComparisonOp(regs.stencil_front_func_func)); + front.action_stencil_fail.Assign(PackStencilOp(regs.stencil_front_op.fail)); + front.action_depth_fail.Assign(PackStencilOp(regs.stencil_front_op.zfail)); + front.action_depth_pass.Assign(PackStencilOp(regs.stencil_front_op.zpass)); + front.test_func.Assign(PackComparisonOp(regs.stencil_front_op.func)); if (regs.stencil_two_side_enable) { - back.action_stencil_fail.Assign(PackStencilOp(regs.stencil_back_op_fail)); - back.action_depth_fail.Assign(PackStencilOp(regs.stencil_back_op_zfail)); - back.action_depth_pass.Assign(PackStencilOp(regs.stencil_back_op_zpass)); - back.test_func.Assign(PackComparisonOp(regs.stencil_back_func_func)); + back.action_stencil_fail.Assign(PackStencilOp(regs.stencil_back_op.fail)); + back.action_depth_fail.Assign(PackStencilOp(regs.stencil_back_op.zfail)); + back.action_depth_pass.Assign(PackStencilOp(regs.stencil_back_op.zpass)); + back.test_func.Assign(PackComparisonOp(regs.stencil_back_op.func)); } else { back.action_stencil_fail.Assign(front.action_stencil_fail); back.action_depth_fail.Assign(front.action_depth_fail); @@ -222,9 +228,9 @@ void FixedPipelineState::DynamicState::Refresh(const Maxwell& regs) { depth_test_enable.Assign(regs.depth_test_enable); front_face.Assign(packed_front_face); depth_test_func.Assign(PackComparisonOp(regs.depth_test_func)); - cull_face.Assign(PackCullFace(regs.cull_face)); - cull_enable.Assign(regs.cull_test_enabled != 0 ? 1 : 0); - std::ranges::transform(regs.vertex_array, vertex_strides.begin(), [](const auto& array) { + cull_face.Assign(PackCullFace(regs.gl_cull_face)); + cull_enable.Assign(regs.gl_cull_test_enabled != 0 ? 1 : 0); + std::ranges::transform(regs.vertex_streams, vertex_strides.begin(), [](const auto& array) { return static_cast<u16>(array.stride.Value()); }); } @@ -251,41 +257,42 @@ Maxwell::ComparisonOp FixedPipelineState::UnpackComparisonOp(u32 packed) noexcep return static_cast<Maxwell::ComparisonOp>(packed + 1); } -u32 FixedPipelineState::PackStencilOp(Maxwell::StencilOp op) noexcept { +u32 FixedPipelineState::PackStencilOp(Maxwell::StencilOp::Op op) noexcept { switch (op) { - case Maxwell::StencilOp::Keep: - case Maxwell::StencilOp::KeepOGL: + case Maxwell::StencilOp::Op::Keep_D3D: + case Maxwell::StencilOp::Op::Keep_GL: return 0; - case Maxwell::StencilOp::Zero: - case Maxwell::StencilOp::ZeroOGL: + case Maxwell::StencilOp::Op::Zero_D3D: + case Maxwell::StencilOp::Op::Zero_GL: return 1; - case Maxwell::StencilOp::Replace: - case Maxwell::StencilOp::ReplaceOGL: + case Maxwell::StencilOp::Op::Replace_D3D: + case Maxwell::StencilOp::Op::Replace_GL: return 2; - case Maxwell::StencilOp::Incr: - case Maxwell::StencilOp::IncrOGL: + case Maxwell::StencilOp::Op::IncrSaturate_D3D: + case Maxwell::StencilOp::Op::IncrSaturate_GL: return 3; - case Maxwell::StencilOp::Decr: - case Maxwell::StencilOp::DecrOGL: + case Maxwell::StencilOp::Op::DecrSaturate_D3D: + case Maxwell::StencilOp::Op::DecrSaturate_GL: return 4; - case Maxwell::StencilOp::Invert: - case Maxwell::StencilOp::InvertOGL: + case Maxwell::StencilOp::Op::Invert_D3D: + case Maxwell::StencilOp::Op::Invert_GL: return 5; - case Maxwell::StencilOp::IncrWrap: - case Maxwell::StencilOp::IncrWrapOGL: + case Maxwell::StencilOp::Op::Incr_D3D: + case Maxwell::StencilOp::Op::Incr_GL: return 6; - case Maxwell::StencilOp::DecrWrap: - case Maxwell::StencilOp::DecrWrapOGL: + case Maxwell::StencilOp::Op::Decr_D3D: + case Maxwell::StencilOp::Op::Decr_GL: return 7; } return 0; } -Maxwell::StencilOp FixedPipelineState::UnpackStencilOp(u32 packed) noexcept { - static constexpr std::array LUT = {Maxwell::StencilOp::Keep, Maxwell::StencilOp::Zero, - Maxwell::StencilOp::Replace, Maxwell::StencilOp::Incr, - Maxwell::StencilOp::Decr, Maxwell::StencilOp::Invert, - Maxwell::StencilOp::IncrWrap, Maxwell::StencilOp::DecrWrap}; +Maxwell::StencilOp::Op FixedPipelineState::UnpackStencilOp(u32 packed) noexcept { + static constexpr std::array LUT = { + Maxwell::StencilOp::Op::Keep_D3D, Maxwell::StencilOp::Op::Zero_D3D, + Maxwell::StencilOp::Op::Replace_D3D, Maxwell::StencilOp::Op::IncrSaturate_D3D, + Maxwell::StencilOp::Op::DecrSaturate_D3D, Maxwell::StencilOp::Op::Invert_D3D, + Maxwell::StencilOp::Op::Incr_D3D, Maxwell::StencilOp::Op::Decr_D3D}; return LUT[packed]; } @@ -318,30 +325,30 @@ Maxwell::PolygonMode FixedPipelineState::UnpackPolygonMode(u32 packed) noexcept return static_cast<Maxwell::PolygonMode>(packed + 0x1B00); } -u32 FixedPipelineState::PackLogicOp(Maxwell::LogicOperation op) noexcept { +u32 FixedPipelineState::PackLogicOp(Maxwell::LogicOp::Op op) noexcept { return static_cast<u32>(op) - 0x1500; } -Maxwell::LogicOperation FixedPipelineState::UnpackLogicOp(u32 packed) noexcept { - return static_cast<Maxwell::LogicOperation>(packed + 0x1500); +Maxwell::LogicOp::Op FixedPipelineState::UnpackLogicOp(u32 packed) noexcept { + return static_cast<Maxwell::LogicOp::Op>(packed + 0x1500); } u32 FixedPipelineState::PackBlendEquation(Maxwell::Blend::Equation equation) noexcept { switch (equation) { - case Maxwell::Blend::Equation::Add: - case Maxwell::Blend::Equation::AddGL: + case Maxwell::Blend::Equation::Add_D3D: + case Maxwell::Blend::Equation::Add_GL: return 0; - case Maxwell::Blend::Equation::Subtract: - case Maxwell::Blend::Equation::SubtractGL: + case Maxwell::Blend::Equation::Subtract_D3D: + case Maxwell::Blend::Equation::Subtract_GL: return 1; - case Maxwell::Blend::Equation::ReverseSubtract: - case Maxwell::Blend::Equation::ReverseSubtractGL: + case Maxwell::Blend::Equation::ReverseSubtract_D3D: + case Maxwell::Blend::Equation::ReverseSubtract_GL: return 2; - case Maxwell::Blend::Equation::Min: - case Maxwell::Blend::Equation::MinGL: + case Maxwell::Blend::Equation::Min_D3D: + case Maxwell::Blend::Equation::Min_GL: return 3; - case Maxwell::Blend::Equation::Max: - case Maxwell::Blend::Equation::MaxGL: + case Maxwell::Blend::Equation::Max_D3D: + case Maxwell::Blend::Equation::Max_GL: return 4; } return 0; @@ -349,97 +356,99 @@ u32 FixedPipelineState::PackBlendEquation(Maxwell::Blend::Equation equation) noe Maxwell::Blend::Equation FixedPipelineState::UnpackBlendEquation(u32 packed) noexcept { static constexpr std::array LUT = { - Maxwell::Blend::Equation::Add, Maxwell::Blend::Equation::Subtract, - Maxwell::Blend::Equation::ReverseSubtract, Maxwell::Blend::Equation::Min, - Maxwell::Blend::Equation::Max}; + Maxwell::Blend::Equation::Add_D3D, Maxwell::Blend::Equation::Subtract_D3D, + Maxwell::Blend::Equation::ReverseSubtract_D3D, Maxwell::Blend::Equation::Min_D3D, + Maxwell::Blend::Equation::Max_D3D}; return LUT[packed]; } u32 FixedPipelineState::PackBlendFactor(Maxwell::Blend::Factor factor) noexcept { switch (factor) { - case Maxwell::Blend::Factor::Zero: - case Maxwell::Blend::Factor::ZeroGL: + case Maxwell::Blend::Factor::Zero_D3D: + case Maxwell::Blend::Factor::Zero_GL: return 0; - case Maxwell::Blend::Factor::One: - case Maxwell::Blend::Factor::OneGL: + case Maxwell::Blend::Factor::One_D3D: + case Maxwell::Blend::Factor::One_GL: return 1; - case Maxwell::Blend::Factor::SourceColor: - case Maxwell::Blend::Factor::SourceColorGL: + case Maxwell::Blend::Factor::SourceColor_D3D: + case Maxwell::Blend::Factor::SourceColor_GL: return 2; - case Maxwell::Blend::Factor::OneMinusSourceColor: - case Maxwell::Blend::Factor::OneMinusSourceColorGL: + case Maxwell::Blend::Factor::OneMinusSourceColor_D3D: + case Maxwell::Blend::Factor::OneMinusSourceColor_GL: return 3; - case Maxwell::Blend::Factor::SourceAlpha: - case Maxwell::Blend::Factor::SourceAlphaGL: + case Maxwell::Blend::Factor::SourceAlpha_D3D: + case Maxwell::Blend::Factor::SourceAlpha_GL: return 4; - case Maxwell::Blend::Factor::OneMinusSourceAlpha: - case Maxwell::Blend::Factor::OneMinusSourceAlphaGL: + case Maxwell::Blend::Factor::OneMinusSourceAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusSourceAlpha_GL: return 5; - case Maxwell::Blend::Factor::DestAlpha: - case Maxwell::Blend::Factor::DestAlphaGL: + case Maxwell::Blend::Factor::DestAlpha_D3D: + case Maxwell::Blend::Factor::DestAlpha_GL: return 6; - case Maxwell::Blend::Factor::OneMinusDestAlpha: - case Maxwell::Blend::Factor::OneMinusDestAlphaGL: + case Maxwell::Blend::Factor::OneMinusDestAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusDestAlpha_GL: return 7; - case Maxwell::Blend::Factor::DestColor: - case Maxwell::Blend::Factor::DestColorGL: + case Maxwell::Blend::Factor::DestColor_D3D: + case Maxwell::Blend::Factor::DestColor_GL: return 8; - case Maxwell::Blend::Factor::OneMinusDestColor: - case Maxwell::Blend::Factor::OneMinusDestColorGL: + case Maxwell::Blend::Factor::OneMinusDestColor_D3D: + case Maxwell::Blend::Factor::OneMinusDestColor_GL: return 9; - case Maxwell::Blend::Factor::SourceAlphaSaturate: - case Maxwell::Blend::Factor::SourceAlphaSaturateGL: + case Maxwell::Blend::Factor::SourceAlphaSaturate_D3D: + case Maxwell::Blend::Factor::SourceAlphaSaturate_GL: return 10; - case Maxwell::Blend::Factor::Source1Color: - case Maxwell::Blend::Factor::Source1ColorGL: + case Maxwell::Blend::Factor::Source1Color_D3D: + case Maxwell::Blend::Factor::Source1Color_GL: return 11; - case Maxwell::Blend::Factor::OneMinusSource1Color: - case Maxwell::Blend::Factor::OneMinusSource1ColorGL: + case Maxwell::Blend::Factor::OneMinusSource1Color_D3D: + case Maxwell::Blend::Factor::OneMinusSource1Color_GL: return 12; - case Maxwell::Blend::Factor::Source1Alpha: - case Maxwell::Blend::Factor::Source1AlphaGL: + case Maxwell::Blend::Factor::Source1Alpha_D3D: + case Maxwell::Blend::Factor::Source1Alpha_GL: return 13; - case Maxwell::Blend::Factor::OneMinusSource1Alpha: - case Maxwell::Blend::Factor::OneMinusSource1AlphaGL: + case Maxwell::Blend::Factor::OneMinusSource1Alpha_D3D: + case Maxwell::Blend::Factor::OneMinusSource1Alpha_GL: return 14; - case Maxwell::Blend::Factor::ConstantColor: - case Maxwell::Blend::Factor::ConstantColorGL: + case Maxwell::Blend::Factor::BlendFactor_D3D: + case Maxwell::Blend::Factor::ConstantColor_GL: return 15; - case Maxwell::Blend::Factor::OneMinusConstantColor: - case Maxwell::Blend::Factor::OneMinusConstantColorGL: + case Maxwell::Blend::Factor::OneMinusBlendFactor_D3D: + case Maxwell::Blend::Factor::OneMinusConstantColor_GL: return 16; - case Maxwell::Blend::Factor::ConstantAlpha: - case Maxwell::Blend::Factor::ConstantAlphaGL: + case Maxwell::Blend::Factor::BothSourceAlpha_D3D: + case Maxwell::Blend::Factor::ConstantAlpha_GL: return 17; - case Maxwell::Blend::Factor::OneMinusConstantAlpha: - case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: + case Maxwell::Blend::Factor::OneMinusBothSourceAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusConstantAlpha_GL: return 18; } + UNIMPLEMENTED_MSG("Unknown blend factor {}", static_cast<u32>(factor)); return 0; } Maxwell::Blend::Factor FixedPipelineState::UnpackBlendFactor(u32 packed) noexcept { static constexpr std::array LUT = { - Maxwell::Blend::Factor::Zero, - Maxwell::Blend::Factor::One, - Maxwell::Blend::Factor::SourceColor, - Maxwell::Blend::Factor::OneMinusSourceColor, - Maxwell::Blend::Factor::SourceAlpha, - Maxwell::Blend::Factor::OneMinusSourceAlpha, - Maxwell::Blend::Factor::DestAlpha, - Maxwell::Blend::Factor::OneMinusDestAlpha, - Maxwell::Blend::Factor::DestColor, - Maxwell::Blend::Factor::OneMinusDestColor, - Maxwell::Blend::Factor::SourceAlphaSaturate, - Maxwell::Blend::Factor::Source1Color, - Maxwell::Blend::Factor::OneMinusSource1Color, - Maxwell::Blend::Factor::Source1Alpha, - Maxwell::Blend::Factor::OneMinusSource1Alpha, - Maxwell::Blend::Factor::ConstantColor, - Maxwell::Blend::Factor::OneMinusConstantColor, - Maxwell::Blend::Factor::ConstantAlpha, - Maxwell::Blend::Factor::OneMinusConstantAlpha, + Maxwell::Blend::Factor::Zero_D3D, + Maxwell::Blend::Factor::One_D3D, + Maxwell::Blend::Factor::SourceColor_D3D, + Maxwell::Blend::Factor::OneMinusSourceColor_D3D, + Maxwell::Blend::Factor::SourceAlpha_D3D, + Maxwell::Blend::Factor::OneMinusSourceAlpha_D3D, + Maxwell::Blend::Factor::DestAlpha_D3D, + Maxwell::Blend::Factor::OneMinusDestAlpha_D3D, + Maxwell::Blend::Factor::DestColor_D3D, + Maxwell::Blend::Factor::OneMinusDestColor_D3D, + Maxwell::Blend::Factor::SourceAlphaSaturate_D3D, + Maxwell::Blend::Factor::Source1Color_D3D, + Maxwell::Blend::Factor::OneMinusSource1Color_D3D, + Maxwell::Blend::Factor::Source1Alpha_D3D, + Maxwell::Blend::Factor::OneMinusSource1Alpha_D3D, + Maxwell::Blend::Factor::BlendFactor_D3D, + Maxwell::Blend::Factor::OneMinusBlendFactor_D3D, + Maxwell::Blend::Factor::BothSourceAlpha_D3D, + Maxwell::Blend::Factor::OneMinusBothSourceAlpha_D3D, }; + ASSERT(packed < LUT.size()); return LUT[packed]; } diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.h b/src/video_core/renderer_vulkan/fixed_pipeline_state.h index 9d60756e5..43441209c 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.h +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.h @@ -21,8 +21,8 @@ struct FixedPipelineState { static u32 PackComparisonOp(Maxwell::ComparisonOp op) noexcept; static Maxwell::ComparisonOp UnpackComparisonOp(u32 packed) noexcept; - static u32 PackStencilOp(Maxwell::StencilOp op) noexcept; - static Maxwell::StencilOp UnpackStencilOp(u32 packed) noexcept; + static u32 PackStencilOp(Maxwell::StencilOp::Op op) noexcept; + static Maxwell::StencilOp::Op UnpackStencilOp(u32 packed) noexcept; static u32 PackCullFace(Maxwell::CullFace cull) noexcept; static Maxwell::CullFace UnpackCullFace(u32 packed) noexcept; @@ -33,8 +33,8 @@ struct FixedPipelineState { static u32 PackPolygonMode(Maxwell::PolygonMode mode) noexcept; static Maxwell::PolygonMode UnpackPolygonMode(u32 packed) noexcept; - static u32 PackLogicOp(Maxwell::LogicOperation op) noexcept; - static Maxwell::LogicOperation UnpackLogicOp(u32 packed) noexcept; + static u32 PackLogicOp(Maxwell::LogicOp::Op op) noexcept; + static Maxwell::LogicOp::Op UnpackLogicOp(u32 packed) noexcept; static u32 PackBlendEquation(Maxwell::Blend::Equation equation) noexcept; static Maxwell::Blend::Equation UnpackBlendEquation(u32 packed) noexcept; @@ -113,15 +113,15 @@ struct FixedPipelineState { BitField<Position + 6, 3, u32> action_depth_pass; BitField<Position + 9, 3, u32> test_func; - Maxwell::StencilOp ActionStencilFail() const noexcept { + Maxwell::StencilOp::Op ActionStencilFail() const noexcept { return UnpackStencilOp(action_stencil_fail); } - Maxwell::StencilOp ActionDepthFail() const noexcept { + Maxwell::StencilOp::Op ActionDepthFail() const noexcept { return UnpackStencilOp(action_depth_fail); } - Maxwell::StencilOp ActionDepthPass() const noexcept { + Maxwell::StencilOp::Op ActionDepthPass() const noexcept { return UnpackStencilOp(action_depth_pass); } @@ -193,7 +193,6 @@ struct FixedPipelineState { BitField<6, 5, u32> depth_format; BitField<11, 1, u32> y_negate; BitField<12, 1, u32> provoking_vertex_last; - BitField<13, 1, u32> conservative_raster_enable; BitField<14, 1, u32> smooth_lines; }; std::array<u8, Maxwell::NumRenderTargets> color_formats; diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp index e7104d377..5c156087b 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp @@ -323,161 +323,182 @@ VkFormat VertexFormat(const Device& device, Maxwell::VertexAttribute::Type type, Maxwell::VertexAttribute::Size size) { const VkFormat format{([&]() { switch (type) { - case Maxwell::VertexAttribute::Type::UnsignedNorm: + case Maxwell::VertexAttribute::Type::UnusedEnumDoNotUseBecauseItWillGoAway: + ASSERT_MSG(false, "Invalid vertex attribute type!"); + break; + case Maxwell::VertexAttribute::Type::UNorm: switch (size) { - case Maxwell::VertexAttribute::Size::Size_8: + case Maxwell::VertexAttribute::Size::Size_R8: + case Maxwell::VertexAttribute::Size::Size_A8: return VK_FORMAT_R8_UNORM; - case Maxwell::VertexAttribute::Size::Size_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8: + case Maxwell::VertexAttribute::Size::Size_G8_R8: return VK_FORMAT_R8G8_UNORM; - case Maxwell::VertexAttribute::Size::Size_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8: return VK_FORMAT_R8G8B8_UNORM; - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8_A8: + case Maxwell::VertexAttribute::Size::Size_X8_B8_G8_R8: return VK_FORMAT_R8G8B8A8_UNORM; - case Maxwell::VertexAttribute::Size::Size_16: + case Maxwell::VertexAttribute::Size::Size_R16: return VK_FORMAT_R16_UNORM; - case Maxwell::VertexAttribute::Size::Size_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: return VK_FORMAT_R16G16_UNORM; - case Maxwell::VertexAttribute::Size::Size_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: return VK_FORMAT_R16G16B16_UNORM; - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return VK_FORMAT_R16G16B16A16_UNORM; - case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + case Maxwell::VertexAttribute::Size::Size_A2_B10_G10_R10: return VK_FORMAT_A2B10G10R10_UNORM_PACK32; default: break; } break; - case Maxwell::VertexAttribute::Type::SignedNorm: + case Maxwell::VertexAttribute::Type::SNorm: switch (size) { - case Maxwell::VertexAttribute::Size::Size_8: + case Maxwell::VertexAttribute::Size::Size_R8: + case Maxwell::VertexAttribute::Size::Size_A8: return VK_FORMAT_R8_SNORM; - case Maxwell::VertexAttribute::Size::Size_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8: + case Maxwell::VertexAttribute::Size::Size_G8_R8: return VK_FORMAT_R8G8_SNORM; - case Maxwell::VertexAttribute::Size::Size_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8: return VK_FORMAT_R8G8B8_SNORM; - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8_A8: + case Maxwell::VertexAttribute::Size::Size_X8_B8_G8_R8: return VK_FORMAT_R8G8B8A8_SNORM; - case Maxwell::VertexAttribute::Size::Size_16: + case Maxwell::VertexAttribute::Size::Size_R16: return VK_FORMAT_R16_SNORM; - case Maxwell::VertexAttribute::Size::Size_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: return VK_FORMAT_R16G16_SNORM; - case Maxwell::VertexAttribute::Size::Size_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: return VK_FORMAT_R16G16B16_SNORM; - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return VK_FORMAT_R16G16B16A16_SNORM; - case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + case Maxwell::VertexAttribute::Size::Size_A2_B10_G10_R10: return VK_FORMAT_A2B10G10R10_SNORM_PACK32; default: break; } break; - case Maxwell::VertexAttribute::Type::UnsignedScaled: + case Maxwell::VertexAttribute::Type::UScaled: switch (size) { - case Maxwell::VertexAttribute::Size::Size_8: + case Maxwell::VertexAttribute::Size::Size_R8: + case Maxwell::VertexAttribute::Size::Size_A8: return VK_FORMAT_R8_USCALED; - case Maxwell::VertexAttribute::Size::Size_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8: + case Maxwell::VertexAttribute::Size::Size_G8_R8: return VK_FORMAT_R8G8_USCALED; - case Maxwell::VertexAttribute::Size::Size_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8: return VK_FORMAT_R8G8B8_USCALED; - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8_A8: + case Maxwell::VertexAttribute::Size::Size_X8_B8_G8_R8: return VK_FORMAT_R8G8B8A8_USCALED; - case Maxwell::VertexAttribute::Size::Size_16: + case Maxwell::VertexAttribute::Size::Size_R16: return VK_FORMAT_R16_USCALED; - case Maxwell::VertexAttribute::Size::Size_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: return VK_FORMAT_R16G16_USCALED; - case Maxwell::VertexAttribute::Size::Size_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: return VK_FORMAT_R16G16B16_USCALED; - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return VK_FORMAT_R16G16B16A16_USCALED; - case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + case Maxwell::VertexAttribute::Size::Size_A2_B10_G10_R10: return VK_FORMAT_A2B10G10R10_USCALED_PACK32; default: break; } break; - case Maxwell::VertexAttribute::Type::SignedScaled: + case Maxwell::VertexAttribute::Type::SScaled: switch (size) { - case Maxwell::VertexAttribute::Size::Size_8: + case Maxwell::VertexAttribute::Size::Size_R8: + case Maxwell::VertexAttribute::Size::Size_A8: return VK_FORMAT_R8_SSCALED; - case Maxwell::VertexAttribute::Size::Size_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8: + case Maxwell::VertexAttribute::Size::Size_G8_R8: return VK_FORMAT_R8G8_SSCALED; - case Maxwell::VertexAttribute::Size::Size_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8: return VK_FORMAT_R8G8B8_SSCALED; - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8_A8: + case Maxwell::VertexAttribute::Size::Size_X8_B8_G8_R8: return VK_FORMAT_R8G8B8A8_SSCALED; - case Maxwell::VertexAttribute::Size::Size_16: + case Maxwell::VertexAttribute::Size::Size_R16: return VK_FORMAT_R16_SSCALED; - case Maxwell::VertexAttribute::Size::Size_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: return VK_FORMAT_R16G16_SSCALED; - case Maxwell::VertexAttribute::Size::Size_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: return VK_FORMAT_R16G16B16_SSCALED; - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return VK_FORMAT_R16G16B16A16_SSCALED; - case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + case Maxwell::VertexAttribute::Size::Size_A2_B10_G10_R10: return VK_FORMAT_A2B10G10R10_SSCALED_PACK32; default: break; } break; - case Maxwell::VertexAttribute::Type::UnsignedInt: + case Maxwell::VertexAttribute::Type::UInt: switch (size) { - case Maxwell::VertexAttribute::Size::Size_8: + case Maxwell::VertexAttribute::Size::Size_R8: + case Maxwell::VertexAttribute::Size::Size_A8: return VK_FORMAT_R8_UINT; - case Maxwell::VertexAttribute::Size::Size_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8: + case Maxwell::VertexAttribute::Size::Size_G8_R8: return VK_FORMAT_R8G8_UINT; - case Maxwell::VertexAttribute::Size::Size_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8: return VK_FORMAT_R8G8B8_UINT; - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8_A8: + case Maxwell::VertexAttribute::Size::Size_X8_B8_G8_R8: return VK_FORMAT_R8G8B8A8_UINT; - case Maxwell::VertexAttribute::Size::Size_16: + case Maxwell::VertexAttribute::Size::Size_R16: return VK_FORMAT_R16_UINT; - case Maxwell::VertexAttribute::Size::Size_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: return VK_FORMAT_R16G16_UINT; - case Maxwell::VertexAttribute::Size::Size_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: return VK_FORMAT_R16G16B16_UINT; - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return VK_FORMAT_R16G16B16A16_UINT; - case Maxwell::VertexAttribute::Size::Size_32: + case Maxwell::VertexAttribute::Size::Size_R32: return VK_FORMAT_R32_UINT; - case Maxwell::VertexAttribute::Size::Size_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32: return VK_FORMAT_R32G32_UINT; - case Maxwell::VertexAttribute::Size::Size_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32: return VK_FORMAT_R32G32B32_UINT; - case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32_A32: return VK_FORMAT_R32G32B32A32_UINT; - case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + case Maxwell::VertexAttribute::Size::Size_A2_B10_G10_R10: return VK_FORMAT_A2B10G10R10_UINT_PACK32; default: break; } break; - case Maxwell::VertexAttribute::Type::SignedInt: + case Maxwell::VertexAttribute::Type::SInt: switch (size) { - case Maxwell::VertexAttribute::Size::Size_8: + case Maxwell::VertexAttribute::Size::Size_R8: + case Maxwell::VertexAttribute::Size::Size_A8: return VK_FORMAT_R8_SINT; - case Maxwell::VertexAttribute::Size::Size_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8: + case Maxwell::VertexAttribute::Size::Size_G8_R8: return VK_FORMAT_R8G8_SINT; - case Maxwell::VertexAttribute::Size::Size_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8: return VK_FORMAT_R8G8B8_SINT; - case Maxwell::VertexAttribute::Size::Size_8_8_8_8: + case Maxwell::VertexAttribute::Size::Size_R8_G8_B8_A8: + case Maxwell::VertexAttribute::Size::Size_X8_B8_G8_R8: return VK_FORMAT_R8G8B8A8_SINT; - case Maxwell::VertexAttribute::Size::Size_16: + case Maxwell::VertexAttribute::Size::Size_R16: return VK_FORMAT_R16_SINT; - case Maxwell::VertexAttribute::Size::Size_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: return VK_FORMAT_R16G16_SINT; - case Maxwell::VertexAttribute::Size::Size_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: return VK_FORMAT_R16G16B16_SINT; - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return VK_FORMAT_R16G16B16A16_SINT; - case Maxwell::VertexAttribute::Size::Size_32: + case Maxwell::VertexAttribute::Size::Size_R32: return VK_FORMAT_R32_SINT; - case Maxwell::VertexAttribute::Size::Size_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32: return VK_FORMAT_R32G32_SINT; - case Maxwell::VertexAttribute::Size::Size_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32: return VK_FORMAT_R32G32B32_SINT; - case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32_A32: return VK_FORMAT_R32G32B32A32_SINT; - case Maxwell::VertexAttribute::Size::Size_10_10_10_2: + case Maxwell::VertexAttribute::Size::Size_A2_B10_G10_R10: return VK_FORMAT_A2B10G10R10_SINT_PACK32; default: break; @@ -485,23 +506,23 @@ VkFormat VertexFormat(const Device& device, Maxwell::VertexAttribute::Type type, break; case Maxwell::VertexAttribute::Type::Float: switch (size) { - case Maxwell::VertexAttribute::Size::Size_16: + case Maxwell::VertexAttribute::Size::Size_R16: return VK_FORMAT_R16_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16: return VK_FORMAT_R16G16_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16: return VK_FORMAT_R16G16B16_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_16_16_16_16: + case Maxwell::VertexAttribute::Size::Size_R16_G16_B16_A16: return VK_FORMAT_R16G16B16A16_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_32: + case Maxwell::VertexAttribute::Size::Size_R32: return VK_FORMAT_R32_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32: return VK_FORMAT_R32G32_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32: return VK_FORMAT_R32G32B32_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_32_32_32_32: + case Maxwell::VertexAttribute::Size::Size_R32_G32_B32_A32: return VK_FORMAT_R32G32B32A32_SFLOAT; - case Maxwell::VertexAttribute::Size::Size_11_11_10: + case Maxwell::VertexAttribute::Size::Size_B10_G11_R11: return VK_FORMAT_B10G11R11_UFLOAT_PACK32; default: break; @@ -521,29 +542,29 @@ VkFormat VertexFormat(const Device& device, Maxwell::VertexAttribute::Type type, VkCompareOp ComparisonOp(Maxwell::ComparisonOp comparison) { switch (comparison) { - case Maxwell::ComparisonOp::Never: - case Maxwell::ComparisonOp::NeverOld: + case Maxwell::ComparisonOp::Never_D3D: + case Maxwell::ComparisonOp::Never_GL: return VK_COMPARE_OP_NEVER; - case Maxwell::ComparisonOp::Less: - case Maxwell::ComparisonOp::LessOld: + case Maxwell::ComparisonOp::Less_D3D: + case Maxwell::ComparisonOp::Less_GL: return VK_COMPARE_OP_LESS; - case Maxwell::ComparisonOp::Equal: - case Maxwell::ComparisonOp::EqualOld: + case Maxwell::ComparisonOp::Equal_D3D: + case Maxwell::ComparisonOp::Equal_GL: return VK_COMPARE_OP_EQUAL; - case Maxwell::ComparisonOp::LessEqual: - case Maxwell::ComparisonOp::LessEqualOld: + case Maxwell::ComparisonOp::LessEqual_D3D: + case Maxwell::ComparisonOp::LessEqual_GL: return VK_COMPARE_OP_LESS_OR_EQUAL; - case Maxwell::ComparisonOp::Greater: - case Maxwell::ComparisonOp::GreaterOld: + case Maxwell::ComparisonOp::Greater_D3D: + case Maxwell::ComparisonOp::Greater_GL: return VK_COMPARE_OP_GREATER; - case Maxwell::ComparisonOp::NotEqual: - case Maxwell::ComparisonOp::NotEqualOld: + case Maxwell::ComparisonOp::NotEqual_D3D: + case Maxwell::ComparisonOp::NotEqual_GL: return VK_COMPARE_OP_NOT_EQUAL; - case Maxwell::ComparisonOp::GreaterEqual: - case Maxwell::ComparisonOp::GreaterEqualOld: + case Maxwell::ComparisonOp::GreaterEqual_D3D: + case Maxwell::ComparisonOp::GreaterEqual_GL: return VK_COMPARE_OP_GREATER_OR_EQUAL; - case Maxwell::ComparisonOp::Always: - case Maxwell::ComparisonOp::AlwaysOld: + case Maxwell::ComparisonOp::Always_D3D: + case Maxwell::ComparisonOp::Always_GL: return VK_COMPARE_OP_ALWAYS; } UNIMPLEMENTED_MSG("Unimplemented comparison op={}", comparison); @@ -563,31 +584,31 @@ VkIndexType IndexFormat(Maxwell::IndexFormat index_format) { return {}; } -VkStencilOp StencilOp(Maxwell::StencilOp stencil_op) { +VkStencilOp StencilOp(Maxwell::StencilOp::Op stencil_op) { switch (stencil_op) { - case Maxwell::StencilOp::Keep: - case Maxwell::StencilOp::KeepOGL: + case Maxwell::StencilOp::Op::Keep_D3D: + case Maxwell::StencilOp::Op::Keep_GL: return VK_STENCIL_OP_KEEP; - case Maxwell::StencilOp::Zero: - case Maxwell::StencilOp::ZeroOGL: + case Maxwell::StencilOp::Op::Zero_D3D: + case Maxwell::StencilOp::Op::Zero_GL: return VK_STENCIL_OP_ZERO; - case Maxwell::StencilOp::Replace: - case Maxwell::StencilOp::ReplaceOGL: + case Maxwell::StencilOp::Op::Replace_D3D: + case Maxwell::StencilOp::Op::Replace_GL: return VK_STENCIL_OP_REPLACE; - case Maxwell::StencilOp::Incr: - case Maxwell::StencilOp::IncrOGL: + case Maxwell::StencilOp::Op::IncrSaturate_D3D: + case Maxwell::StencilOp::Op::IncrSaturate_GL: return VK_STENCIL_OP_INCREMENT_AND_CLAMP; - case Maxwell::StencilOp::Decr: - case Maxwell::StencilOp::DecrOGL: + case Maxwell::StencilOp::Op::DecrSaturate_D3D: + case Maxwell::StencilOp::Op::DecrSaturate_GL: return VK_STENCIL_OP_DECREMENT_AND_CLAMP; - case Maxwell::StencilOp::Invert: - case Maxwell::StencilOp::InvertOGL: + case Maxwell::StencilOp::Op::Invert_D3D: + case Maxwell::StencilOp::Op::Invert_GL: return VK_STENCIL_OP_INVERT; - case Maxwell::StencilOp::IncrWrap: - case Maxwell::StencilOp::IncrWrapOGL: + case Maxwell::StencilOp::Op::Incr_D3D: + case Maxwell::StencilOp::Op::Incr_GL: return VK_STENCIL_OP_INCREMENT_AND_WRAP; - case Maxwell::StencilOp::DecrWrap: - case Maxwell::StencilOp::DecrWrapOGL: + case Maxwell::StencilOp::Op::Decr_D3D: + case Maxwell::StencilOp::Op::Decr_GL: return VK_STENCIL_OP_DECREMENT_AND_WRAP; } UNIMPLEMENTED_MSG("Unimplemented stencil op={}", stencil_op); @@ -596,20 +617,20 @@ VkStencilOp StencilOp(Maxwell::StencilOp stencil_op) { VkBlendOp BlendEquation(Maxwell::Blend::Equation equation) { switch (equation) { - case Maxwell::Blend::Equation::Add: - case Maxwell::Blend::Equation::AddGL: + case Maxwell::Blend::Equation::Add_D3D: + case Maxwell::Blend::Equation::Add_GL: return VK_BLEND_OP_ADD; - case Maxwell::Blend::Equation::Subtract: - case Maxwell::Blend::Equation::SubtractGL: + case Maxwell::Blend::Equation::Subtract_D3D: + case Maxwell::Blend::Equation::Subtract_GL: return VK_BLEND_OP_SUBTRACT; - case Maxwell::Blend::Equation::ReverseSubtract: - case Maxwell::Blend::Equation::ReverseSubtractGL: + case Maxwell::Blend::Equation::ReverseSubtract_D3D: + case Maxwell::Blend::Equation::ReverseSubtract_GL: return VK_BLEND_OP_REVERSE_SUBTRACT; - case Maxwell::Blend::Equation::Min: - case Maxwell::Blend::Equation::MinGL: + case Maxwell::Blend::Equation::Min_D3D: + case Maxwell::Blend::Equation::Min_GL: return VK_BLEND_OP_MIN; - case Maxwell::Blend::Equation::Max: - case Maxwell::Blend::Equation::MaxGL: + case Maxwell::Blend::Equation::Max_D3D: + case Maxwell::Blend::Equation::Max_GL: return VK_BLEND_OP_MAX; } UNIMPLEMENTED_MSG("Unimplemented blend equation={}", equation); @@ -618,62 +639,62 @@ VkBlendOp BlendEquation(Maxwell::Blend::Equation equation) { VkBlendFactor BlendFactor(Maxwell::Blend::Factor factor) { switch (factor) { - case Maxwell::Blend::Factor::Zero: - case Maxwell::Blend::Factor::ZeroGL: + case Maxwell::Blend::Factor::Zero_D3D: + case Maxwell::Blend::Factor::Zero_GL: return VK_BLEND_FACTOR_ZERO; - case Maxwell::Blend::Factor::One: - case Maxwell::Blend::Factor::OneGL: + case Maxwell::Blend::Factor::One_D3D: + case Maxwell::Blend::Factor::One_GL: return VK_BLEND_FACTOR_ONE; - case Maxwell::Blend::Factor::SourceColor: - case Maxwell::Blend::Factor::SourceColorGL: + case Maxwell::Blend::Factor::SourceColor_D3D: + case Maxwell::Blend::Factor::SourceColor_GL: return VK_BLEND_FACTOR_SRC_COLOR; - case Maxwell::Blend::Factor::OneMinusSourceColor: - case Maxwell::Blend::Factor::OneMinusSourceColorGL: + case Maxwell::Blend::Factor::OneMinusSourceColor_D3D: + case Maxwell::Blend::Factor::OneMinusSourceColor_GL: return VK_BLEND_FACTOR_ONE_MINUS_SRC_COLOR; - case Maxwell::Blend::Factor::SourceAlpha: - case Maxwell::Blend::Factor::SourceAlphaGL: + case Maxwell::Blend::Factor::SourceAlpha_D3D: + case Maxwell::Blend::Factor::SourceAlpha_GL: return VK_BLEND_FACTOR_SRC_ALPHA; - case Maxwell::Blend::Factor::OneMinusSourceAlpha: - case Maxwell::Blend::Factor::OneMinusSourceAlphaGL: + case Maxwell::Blend::Factor::OneMinusSourceAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusSourceAlpha_GL: return VK_BLEND_FACTOR_ONE_MINUS_SRC_ALPHA; - case Maxwell::Blend::Factor::DestAlpha: - case Maxwell::Blend::Factor::DestAlphaGL: + case Maxwell::Blend::Factor::DestAlpha_D3D: + case Maxwell::Blend::Factor::DestAlpha_GL: return VK_BLEND_FACTOR_DST_ALPHA; - case Maxwell::Blend::Factor::OneMinusDestAlpha: - case Maxwell::Blend::Factor::OneMinusDestAlphaGL: + case Maxwell::Blend::Factor::OneMinusDestAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusDestAlpha_GL: return VK_BLEND_FACTOR_ONE_MINUS_DST_ALPHA; - case Maxwell::Blend::Factor::DestColor: - case Maxwell::Blend::Factor::DestColorGL: + case Maxwell::Blend::Factor::DestColor_D3D: + case Maxwell::Blend::Factor::DestColor_GL: return VK_BLEND_FACTOR_DST_COLOR; - case Maxwell::Blend::Factor::OneMinusDestColor: - case Maxwell::Blend::Factor::OneMinusDestColorGL: + case Maxwell::Blend::Factor::OneMinusDestColor_D3D: + case Maxwell::Blend::Factor::OneMinusDestColor_GL: return VK_BLEND_FACTOR_ONE_MINUS_DST_COLOR; - case Maxwell::Blend::Factor::SourceAlphaSaturate: - case Maxwell::Blend::Factor::SourceAlphaSaturateGL: + case Maxwell::Blend::Factor::SourceAlphaSaturate_D3D: + case Maxwell::Blend::Factor::SourceAlphaSaturate_GL: return VK_BLEND_FACTOR_SRC_ALPHA_SATURATE; - case Maxwell::Blend::Factor::Source1Color: - case Maxwell::Blend::Factor::Source1ColorGL: + case Maxwell::Blend::Factor::Source1Color_D3D: + case Maxwell::Blend::Factor::Source1Color_GL: return VK_BLEND_FACTOR_SRC1_COLOR; - case Maxwell::Blend::Factor::OneMinusSource1Color: - case Maxwell::Blend::Factor::OneMinusSource1ColorGL: + case Maxwell::Blend::Factor::OneMinusSource1Color_D3D: + case Maxwell::Blend::Factor::OneMinusSource1Color_GL: return VK_BLEND_FACTOR_ONE_MINUS_SRC1_COLOR; - case Maxwell::Blend::Factor::Source1Alpha: - case Maxwell::Blend::Factor::Source1AlphaGL: + case Maxwell::Blend::Factor::Source1Alpha_D3D: + case Maxwell::Blend::Factor::Source1Alpha_GL: return VK_BLEND_FACTOR_SRC1_ALPHA; - case Maxwell::Blend::Factor::OneMinusSource1Alpha: - case Maxwell::Blend::Factor::OneMinusSource1AlphaGL: + case Maxwell::Blend::Factor::OneMinusSource1Alpha_D3D: + case Maxwell::Blend::Factor::OneMinusSource1Alpha_GL: return VK_BLEND_FACTOR_ONE_MINUS_SRC1_ALPHA; - case Maxwell::Blend::Factor::ConstantColor: - case Maxwell::Blend::Factor::ConstantColorGL: + case Maxwell::Blend::Factor::BlendFactor_D3D: + case Maxwell::Blend::Factor::ConstantColor_GL: return VK_BLEND_FACTOR_CONSTANT_COLOR; - case Maxwell::Blend::Factor::OneMinusConstantColor: - case Maxwell::Blend::Factor::OneMinusConstantColorGL: + case Maxwell::Blend::Factor::OneMinusBlendFactor_D3D: + case Maxwell::Blend::Factor::OneMinusConstantColor_GL: return VK_BLEND_FACTOR_ONE_MINUS_CONSTANT_COLOR; - case Maxwell::Blend::Factor::ConstantAlpha: - case Maxwell::Blend::Factor::ConstantAlphaGL: + case Maxwell::Blend::Factor::BothSourceAlpha_D3D: + case Maxwell::Blend::Factor::ConstantAlpha_GL: return VK_BLEND_FACTOR_CONSTANT_ALPHA; - case Maxwell::Blend::Factor::OneMinusConstantAlpha: - case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: + case Maxwell::Blend::Factor::OneMinusBothSourceAlpha_D3D: + case Maxwell::Blend::Factor::OneMinusConstantAlpha_GL: return VK_BLEND_FACTOR_ONE_MINUS_CONSTANT_ALPHA; } UNIMPLEMENTED_MSG("Unimplemented blend factor={}", factor); diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.h b/src/video_core/renderer_vulkan/maxwell_to_vk.h index 356d46292..6f65502d6 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.h +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.h @@ -55,7 +55,7 @@ VkCompareOp ComparisonOp(Maxwell::ComparisonOp comparison); VkIndexType IndexFormat(Maxwell::IndexFormat index_format); -VkStencilOp StencilOp(Maxwell::StencilOp stencil_op); +VkStencilOp StencilOp(Maxwell::StencilOp::Op stencil_op); VkBlendOp BlendEquation(Maxwell::Blend::Equation equation); diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp index f47786f48..c3f66c8a3 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp @@ -288,7 +288,7 @@ void GraphicsPipeline::ConfigureImpl(bool is_indexed) { buffer_cache.SetUniformBuffersState(enabled_uniform_buffer_masks, &uniform_buffer_sizes); const auto& regs{maxwell3d->regs}; - const bool via_header_index{regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex}; + const bool via_header_index{regs.sampler_binding == Maxwell::SamplerBinding::ViaHeaderBinding}; const auto config_stage{[&](size_t stage) LAMBDA_FORCEINLINE { const Shader::Info& info{stage_infos[stage]}; buffer_cache.UnbindGraphicsStorageBuffers(stage); @@ -664,15 +664,6 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) { .lineStippleFactor = 0, .lineStipplePattern = 0, }; - VkPipelineRasterizationConservativeStateCreateInfoEXT conservative_raster{ - .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_CONSERVATIVE_STATE_CREATE_INFO_EXT, - .pNext = nullptr, - .flags = 0, - .conservativeRasterizationMode = key.state.conservative_raster_enable != 0 - ? VK_CONSERVATIVE_RASTERIZATION_MODE_OVERESTIMATE_EXT - : VK_CONSERVATIVE_RASTERIZATION_MODE_DISABLED_EXT, - .extraPrimitiveOverestimationSize = 0.0f, - }; VkPipelineRasterizationProvokingVertexStateCreateInfoEXT provoking_vertex{ .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_PROVOKING_VERTEX_STATE_CREATE_INFO_EXT, .pNext = nullptr, @@ -683,9 +674,6 @@ void GraphicsPipeline::MakePipeline(VkRenderPass render_pass) { if (IsLine(input_assembly_topology) && device.IsExtLineRasterizationSupported()) { line_state.pNext = std::exchange(rasterization_ci.pNext, &line_state); } - if (device.IsExtConservativeRasterizationSupported()) { - conservative_raster.pNext = std::exchange(rasterization_ci.pNext, &conservative_raster); - } if (device.IsExtProvokingVertexSupported()) { provoking_vertex.pNext = std::exchange(rasterization_ci.pNext, &provoking_vertex); } diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp index 732e7b6f2..20f1d6584 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp @@ -62,29 +62,29 @@ auto MakeSpan(Container& container) { Shader::CompareFunction MaxwellToCompareFunction(Maxwell::ComparisonOp comparison) { switch (comparison) { - case Maxwell::ComparisonOp::Never: - case Maxwell::ComparisonOp::NeverOld: + case Maxwell::ComparisonOp::Never_D3D: + case Maxwell::ComparisonOp::Never_GL: return Shader::CompareFunction::Never; - case Maxwell::ComparisonOp::Less: - case Maxwell::ComparisonOp::LessOld: + case Maxwell::ComparisonOp::Less_D3D: + case Maxwell::ComparisonOp::Less_GL: return Shader::CompareFunction::Less; - case Maxwell::ComparisonOp::Equal: - case Maxwell::ComparisonOp::EqualOld: + case Maxwell::ComparisonOp::Equal_D3D: + case Maxwell::ComparisonOp::Equal_GL: return Shader::CompareFunction::Equal; - case Maxwell::ComparisonOp::LessEqual: - case Maxwell::ComparisonOp::LessEqualOld: + case Maxwell::ComparisonOp::LessEqual_D3D: + case Maxwell::ComparisonOp::LessEqual_GL: return Shader::CompareFunction::LessThanEqual; - case Maxwell::ComparisonOp::Greater: - case Maxwell::ComparisonOp::GreaterOld: + case Maxwell::ComparisonOp::Greater_D3D: + case Maxwell::ComparisonOp::Greater_GL: return Shader::CompareFunction::Greater; - case Maxwell::ComparisonOp::NotEqual: - case Maxwell::ComparisonOp::NotEqualOld: + case Maxwell::ComparisonOp::NotEqual_D3D: + case Maxwell::ComparisonOp::NotEqual_GL: return Shader::CompareFunction::NotEqual; - case Maxwell::ComparisonOp::GreaterEqual: - case Maxwell::ComparisonOp::GreaterEqualOld: + case Maxwell::ComparisonOp::GreaterEqual_D3D: + case Maxwell::ComparisonOp::GreaterEqual_GL: return Shader::CompareFunction::GreaterThanEqual; - case Maxwell::ComparisonOp::Always: - case Maxwell::ComparisonOp::AlwaysOld: + case Maxwell::ComparisonOp::Always_D3D: + case Maxwell::ComparisonOp::Always_GL: return Shader::CompareFunction::Always; } UNIMPLEMENTED_MSG("Unimplemented comparison op={}", comparison); @@ -96,15 +96,18 @@ Shader::AttributeType CastAttributeType(const FixedPipelineState::VertexAttribut return Shader::AttributeType::Disabled; } switch (attr.Type()) { - case Maxwell::VertexAttribute::Type::SignedNorm: - case Maxwell::VertexAttribute::Type::UnsignedNorm: - case Maxwell::VertexAttribute::Type::UnsignedScaled: - case Maxwell::VertexAttribute::Type::SignedScaled: + case Maxwell::VertexAttribute::Type::UnusedEnumDoNotUseBecauseItWillGoAway: + ASSERT_MSG(false, "Invalid vertex attribute type!"); + return Shader::AttributeType::Disabled; + case Maxwell::VertexAttribute::Type::SNorm: + case Maxwell::VertexAttribute::Type::UNorm: + case Maxwell::VertexAttribute::Type::UScaled: + case Maxwell::VertexAttribute::Type::SScaled: case Maxwell::VertexAttribute::Type::Float: return Shader::AttributeType::Float; - case Maxwell::VertexAttribute::Type::SignedInt: + case Maxwell::VertexAttribute::Type::SInt: return Shader::AttributeType::SignedInt; - case Maxwell::VertexAttribute::Type::UnsignedInt: + case Maxwell::VertexAttribute::Type::UInt: return Shader::AttributeType::UnsignedInt; } return Shader::AttributeType::Float; @@ -162,16 +165,14 @@ Shader::RuntimeInfo MakeRuntimeInfo(std::span<const Shader::IR::Program> program } break; case Shader::Stage::TessellationEval: - // We have to flip tessellation clockwise for some reason... - info.tess_clockwise = key.state.tessellation_clockwise == 0; info.tess_primitive = [&key] { const u32 raw{key.state.tessellation_primitive.Value()}; - switch (static_cast<Maxwell::TessellationPrimitive>(raw)) { - case Maxwell::TessellationPrimitive::Isolines: + switch (static_cast<Maxwell::Tessellation::DomainType>(raw)) { + case Maxwell::Tessellation::DomainType::Isolines: return Shader::TessPrimitive::Isolines; - case Maxwell::TessellationPrimitive::Triangles: + case Maxwell::Tessellation::DomainType::Triangles: return Shader::TessPrimitive::Triangles; - case Maxwell::TessellationPrimitive::Quads: + case Maxwell::Tessellation::DomainType::Quads: return Shader::TessPrimitive::Quads; } ASSERT(false); @@ -179,12 +180,12 @@ Shader::RuntimeInfo MakeRuntimeInfo(std::span<const Shader::IR::Program> program }(); info.tess_spacing = [&] { const u32 raw{key.state.tessellation_spacing}; - switch (static_cast<Maxwell::TessellationSpacing>(raw)) { - case Maxwell::TessellationSpacing::Equal: + switch (static_cast<Maxwell::Tessellation::Spacing>(raw)) { + case Maxwell::Tessellation::Spacing::Integer: return Shader::TessSpacing::Equal; - case Maxwell::TessellationSpacing::FractionalOdd: + case Maxwell::Tessellation::Spacing::FractionalOdd: return Shader::TessSpacing::FractionalOdd; - case Maxwell::TessellationSpacing::FractionalEven: + case Maxwell::Tessellation::Spacing::FractionalEven: return Shader::TessSpacing::FractionalEven; } ASSERT(false); @@ -490,7 +491,7 @@ GraphicsPipeline* PipelineCache::BuiltPipeline(GraphicsPipeline* pipeline) const // If games are using a small index count, we can assume these are full screen quads. // Usually these shaders are only used once for building textures so we can assume they // can't be built async - if (maxwell3d->regs.index_array.count <= 6 || maxwell3d->regs.vertex_buffer.count <= 6) { + if (maxwell3d->regs.index_buffer.count <= 6 || maxwell3d->regs.vertex_buffer.count <= 6) { return pipeline; } return nullptr; diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index acfd5da7d..47dfb45a1 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -70,7 +70,7 @@ VkViewport GetViewportState(const Device& device, const Maxwell& regs, size_t in const float width = conv(src.scale_x * 2.0f); float y = conv(src.translate_y - src.scale_y); float height = conv(src.scale_y * 2.0f); - bool y_negate = regs.screen_y_control.y_negate; + bool y_negate = regs.window_origin.mode != Maxwell::WindowOrigin::Mode::UpperLeft; if (!device.IsNvViewportSwizzleSupported()) { y_negate = y_negate != (src.swizzle.y == Maxwell::ViewportSwizzle::NegativeY); @@ -130,11 +130,11 @@ VkRect2D GetScissorState(const Maxwell& regs, size_t index, u32 up_scale = 1, u3 DrawParams MakeDrawParams(const Maxwell& regs, u32 num_instances, bool is_instanced, bool is_indexed) { DrawParams params{ - .base_instance = regs.vb_base_instance, + .base_instance = regs.global_base_instance_index, .num_instances = is_instanced ? num_instances : 1, - .base_vertex = is_indexed ? regs.vb_element_base : regs.vertex_buffer.first, - .num_vertices = is_indexed ? regs.index_array.count : regs.vertex_buffer.count, - .first_index = is_indexed ? regs.index_array.first : 0, + .base_vertex = is_indexed ? regs.global_base_vertex_index : regs.vertex_buffer.first, + .num_vertices = is_indexed ? regs.index_buffer.count : regs.vertex_buffer.count, + .first_index = is_indexed ? regs.index_buffer.first : 0, .is_indexed = is_indexed, }; if (regs.draw.topology == Maxwell::PrimitiveTopology::Quads) { @@ -225,10 +225,10 @@ void RasterizerVulkan::Clear() { query_cache.UpdateCounters(); auto& regs = maxwell3d->regs; - const bool use_color = regs.clear_buffers.R || regs.clear_buffers.G || regs.clear_buffers.B || - regs.clear_buffers.A; - const bool use_depth = regs.clear_buffers.Z; - const bool use_stencil = regs.clear_buffers.S; + const bool use_color = regs.clear_surface.R || regs.clear_surface.G || regs.clear_surface.B || + regs.clear_surface.A; + const bool use_depth = regs.clear_surface.Z; + const bool use_stencil = regs.clear_surface.S; if (!use_color && !use_depth && !use_stencil) { return; } @@ -254,9 +254,9 @@ void RasterizerVulkan::Clear() { default_scissor.extent.height = std::numeric_limits<s32>::max(); VkClearRect clear_rect{ - .rect = regs.clear_flags.scissor ? GetScissorState(regs, 0, up_scale, down_shift) - : default_scissor, - .baseArrayLayer = regs.clear_buffers.layer, + .rect = regs.clear_control.use_scissor ? GetScissorState(regs, 0, up_scale, down_shift) + : default_scissor, + .baseArrayLayer = regs.clear_surface.layer, .layerCount = 1, }; if (clear_rect.rect.extent.width == 0 || clear_rect.rect.extent.height == 0) { @@ -267,7 +267,7 @@ void RasterizerVulkan::Clear() { .height = std::min(clear_rect.rect.extent.height, render_area.height), }; - const u32 color_attachment = regs.clear_buffers.RT; + const u32 color_attachment = regs.clear_surface.RT; if (use_color && framebuffer->HasAspectColorBit(color_attachment)) { VkClearValue clear_value; bool is_integer = false; @@ -289,7 +289,8 @@ void RasterizerVulkan::Clear() { break; } if (!is_integer) { - std::memcpy(clear_value.color.float32, regs.clear_color, sizeof(regs.clear_color)); + std::memcpy(clear_value.color.float32, regs.clear_color.data(), + regs.clear_color.size() * sizeof(f32)); } else if (!is_signed) { for (size_t i = 0; i < 4; i++) { clear_value.color.uint32[i] = static_cast<u32>( @@ -648,23 +649,23 @@ void RasterizerVulkan::UpdateDynamicStates() { void RasterizerVulkan::BeginTransformFeedback() { const auto& regs = maxwell3d->regs; - if (regs.tfb_enabled == 0) { + if (regs.transform_feedback_enabled == 0) { return; } if (!device.IsExtTransformFeedbackSupported()) { LOG_ERROR(Render_Vulkan, "Transform feedbacks used but not supported"); return; } - UNIMPLEMENTED_IF(regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::TesselationControl) || - regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::TesselationEval) || - regs.IsShaderConfigEnabled(Maxwell::ShaderProgram::Geometry)); + UNIMPLEMENTED_IF(regs.IsShaderConfigEnabled(Maxwell::ShaderType::TessellationInit) || + regs.IsShaderConfigEnabled(Maxwell::ShaderType::Tessellation) || + regs.IsShaderConfigEnabled(Maxwell::ShaderType::Geometry)); scheduler.Record( [](vk::CommandBuffer cmdbuf) { cmdbuf.BeginTransformFeedbackEXT(0, 0, nullptr, nullptr); }); } void RasterizerVulkan::EndTransformFeedback() { const auto& regs = maxwell3d->regs; - if (regs.tfb_enabled == 0) { + if (regs.transform_feedback_enabled == 0) { return; } if (!device.IsExtTransformFeedbackSupported()) { @@ -728,11 +729,11 @@ void RasterizerVulkan::UpdateDepthBias(Tegra::Engines::Maxwell3D::Regs& regs) { if (!state_tracker.TouchDepthBias()) { return; } - float units = regs.polygon_offset_units / 2.0f; - const bool is_d24 = regs.zeta.format == Tegra::DepthFormat::S8_UINT_Z24_UNORM || - regs.zeta.format == Tegra::DepthFormat::D24X8_UNORM || - regs.zeta.format == Tegra::DepthFormat::D24S8_UNORM || - regs.zeta.format == Tegra::DepthFormat::D24C8_UNORM; + float units = regs.depth_bias / 2.0f; + const bool is_d24 = regs.zeta.format == Tegra::DepthFormat::Z24_UNORM_S8_UINT || + regs.zeta.format == Tegra::DepthFormat::X8Z24_UNORM || + regs.zeta.format == Tegra::DepthFormat::S8Z24_UNORM || + regs.zeta.format == Tegra::DepthFormat::V8Z24_UNORM; if (is_d24 && !device.SupportsD24DepthBuffer()) { // the base formulas can be obtained from here: // https://docs.microsoft.com/en-us/windows/win32/direct3d11/d3d10-graphics-programming-guide-output-merger-stage-depth-bias @@ -740,8 +741,8 @@ void RasterizerVulkan::UpdateDepthBias(Tegra::Engines::Maxwell3D::Regs& regs) { static_cast<double>(1ULL << (32 - 24)) / (static_cast<double>(0x1.ep+127)); units = static_cast<float>(static_cast<double>(units) * rescale_factor); } - scheduler.Record([constant = units, clamp = regs.polygon_offset_clamp, - factor = regs.polygon_offset_factor](vk::CommandBuffer cmdbuf) { + scheduler.Record([constant = units, clamp = regs.depth_bias_clamp, + factor = regs.slope_scale_depth_bias](vk::CommandBuffer cmdbuf) { cmdbuf.SetDepthBias(constant, clamp, factor); }); } @@ -771,8 +772,8 @@ void RasterizerVulkan::UpdateStencilFaces(Tegra::Engines::Maxwell3D::Regs& regs) if (regs.stencil_two_side_enable) { // Separate values per face scheduler.Record( - [front_ref = regs.stencil_front_func_ref, front_write_mask = regs.stencil_front_mask, - front_test_mask = regs.stencil_front_func_mask, back_ref = regs.stencil_back_func_ref, + [front_ref = regs.stencil_front_ref, front_write_mask = regs.stencil_front_mask, + front_test_mask = regs.stencil_front_func_mask, back_ref = regs.stencil_back_ref, back_write_mask = regs.stencil_back_mask, back_test_mask = regs.stencil_back_func_mask](vk::CommandBuffer cmdbuf) { // Front face @@ -787,7 +788,7 @@ void RasterizerVulkan::UpdateStencilFaces(Tegra::Engines::Maxwell3D::Regs& regs) }); } else { // Front face defines both faces - scheduler.Record([ref = regs.stencil_front_func_ref, write_mask = regs.stencil_front_mask, + scheduler.Record([ref = regs.stencil_front_ref, write_mask = regs.stencil_front_mask, test_mask = regs.stencil_front_func_mask](vk::CommandBuffer cmdbuf) { cmdbuf.SetStencilReference(VK_STENCIL_FACE_FRONT_AND_BACK, ref); cmdbuf.SetStencilWriteMask(VK_STENCIL_FACE_FRONT_AND_BACK, write_mask); @@ -800,7 +801,8 @@ void RasterizerVulkan::UpdateLineWidth(Tegra::Engines::Maxwell3D::Regs& regs) { if (!state_tracker.TouchLineWidth()) { return; } - const float width = regs.line_smooth_enable ? regs.line_width_smooth : regs.line_width_aliased; + const float width = + regs.line_anti_alias_enable ? regs.line_width_smooth : regs.line_width_aliased; scheduler.Record([width](vk::CommandBuffer cmdbuf) { cmdbuf.SetLineWidth(width); }); } @@ -808,10 +810,10 @@ void RasterizerVulkan::UpdateCullMode(Tegra::Engines::Maxwell3D::Regs& regs) { if (!state_tracker.TouchCullMode()) { return; } - scheduler.Record( - [enabled = regs.cull_test_enabled, cull_face = regs.cull_face](vk::CommandBuffer cmdbuf) { - cmdbuf.SetCullModeEXT(enabled ? MaxwellToVK::CullFace(cull_face) : VK_CULL_MODE_NONE); - }); + scheduler.Record([enabled = regs.gl_cull_test_enabled, + cull_face = regs.gl_cull_face](vk::CommandBuffer cmdbuf) { + cmdbuf.SetCullModeEXT(enabled ? MaxwellToVK::CullFace(cull_face) : VK_CULL_MODE_NONE); + }); } void RasterizerVulkan::UpdateDepthBoundsTestEnable(Tegra::Engines::Maxwell3D::Regs& regs) { @@ -860,8 +862,8 @@ void RasterizerVulkan::UpdateFrontFace(Tegra::Engines::Maxwell3D::Regs& regs) { return; } - VkFrontFace front_face = MaxwellToVK::FrontFace(regs.front_face); - if (regs.screen_y_control.triangle_rast_flip != 0) { + VkFrontFace front_face = MaxwellToVK::FrontFace(regs.gl_front_face); + if (regs.window_origin.flip_y != 0) { front_face = front_face == VK_FRONT_FACE_CLOCKWISE ? VK_FRONT_FACE_COUNTER_CLOCKWISE : VK_FRONT_FACE_CLOCKWISE; } @@ -873,16 +875,16 @@ void RasterizerVulkan::UpdateStencilOp(Tegra::Engines::Maxwell3D::Regs& regs) { if (!state_tracker.TouchStencilOp()) { return; } - const Maxwell::StencilOp fail = regs.stencil_front_op_fail; - const Maxwell::StencilOp zfail = regs.stencil_front_op_zfail; - const Maxwell::StencilOp zpass = regs.stencil_front_op_zpass; - const Maxwell::ComparisonOp compare = regs.stencil_front_func_func; + const Maxwell::StencilOp::Op fail = regs.stencil_front_op.fail; + const Maxwell::StencilOp::Op zfail = regs.stencil_front_op.zfail; + const Maxwell::StencilOp::Op zpass = regs.stencil_front_op.zpass; + const Maxwell::ComparisonOp compare = regs.stencil_front_op.func; if (regs.stencil_two_side_enable) { // Separate stencil op per face - const Maxwell::StencilOp back_fail = regs.stencil_back_op_fail; - const Maxwell::StencilOp back_zfail = regs.stencil_back_op_zfail; - const Maxwell::StencilOp back_zpass = regs.stencil_back_op_zpass; - const Maxwell::ComparisonOp back_compare = regs.stencil_back_func_func; + const Maxwell::StencilOp::Op back_fail = regs.stencil_back_op.fail; + const Maxwell::StencilOp::Op back_zfail = regs.stencil_back_op.zfail; + const Maxwell::StencilOp::Op back_zpass = regs.stencil_back_op.zpass; + const Maxwell::ComparisonOp back_compare = regs.stencil_back_op.func; scheduler.Record([fail, zfail, zpass, compare, back_fail, back_zfail, back_zpass, back_compare](vk::CommandBuffer cmdbuf) { cmdbuf.SetStencilOpEXT(VK_STENCIL_FACE_FRONT_BIT, MaxwellToVK::StencilOp(fail), @@ -954,15 +956,15 @@ void RasterizerVulkan::UpdateVertexInput(Tegra::Engines::Maxwell3D::Regs& regs) dirty[Dirty::VertexBinding0 + index] = false; const u32 binding{static_cast<u32>(index)}; - const auto& input_binding{regs.vertex_array[binding]}; - const bool is_instanced{regs.instanced_arrays.IsInstancingEnabled(binding)}; + const auto& input_binding{regs.vertex_streams[binding]}; + const bool is_instanced{regs.vertex_stream_instances.IsInstancingEnabled(binding)}; bindings.push_back({ .sType = VK_STRUCTURE_TYPE_VERTEX_INPUT_BINDING_DESCRIPTION_2_EXT, .pNext = nullptr, .binding = binding, .stride = input_binding.stride, .inputRate = is_instanced ? VK_VERTEX_INPUT_RATE_INSTANCE : VK_VERTEX_INPUT_RATE_VERTEX, - .divisor = is_instanced ? input_binding.divisor : 1, + .divisor = is_instanced ? input_binding.frequency : 1, }); } scheduler.Record([bindings, attributes](vk::CommandBuffer cmdbuf) { diff --git a/src/video_core/renderer_vulkan/vk_scheduler.cpp b/src/video_core/renderer_vulkan/vk_scheduler.cpp index d96720b80..7934f2a51 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.cpp +++ b/src/video_core/renderer_vulkan/vk_scheduler.cpp @@ -139,6 +139,7 @@ void Scheduler::WorkerThread(std::stop_token stop_token) { Common::SetCurrentThreadName("VulkanWorker"); do { std::unique_ptr<CommandChunk> work; + bool has_submit{false}; { std::unique_lock lock{work_mutex}; if (work_queue.empty()) { @@ -150,9 +151,10 @@ void Scheduler::WorkerThread(std::stop_token stop_token) { } work = std::move(work_queue.front()); work_queue.pop(); + + has_submit = work->HasSubmit(); + work->ExecuteAll(current_cmdbuf); } - const bool has_submit = work->HasSubmit(); - work->ExecuteAll(current_cmdbuf); if (has_submit) { AllocateWorkerCommandBuffer(); } diff --git a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp index 7fb256953..06f68d09a 100644 --- a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp +++ b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp @@ -26,39 +26,20 @@ using namespace Common::Literals; constexpr VkDeviceSize MAX_ALIGNMENT = 256; // Maximum size to put elements in the stream buffer constexpr VkDeviceSize MAX_STREAM_BUFFER_REQUEST_SIZE = 8_MiB; +// Stream buffer size in bytes +constexpr VkDeviceSize STREAM_BUFFER_SIZE = 128_MiB; +constexpr VkDeviceSize REGION_SIZE = STREAM_BUFFER_SIZE / StagingBufferPool::NUM_SYNCS; constexpr VkMemoryPropertyFlags HOST_FLAGS = VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT; constexpr VkMemoryPropertyFlags STREAM_FLAGS = VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT | HOST_FLAGS; -static bool IsStreamHeap(VkMemoryHeap heap, size_t staging_buffer_size) noexcept { - return staging_buffer_size < (heap.size * 2) / 3; -} - -static bool HasLargeDeviceLocalHostVisibleMemory(const VkPhysicalDeviceMemoryProperties& props) { - const auto flags{VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT | VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT}; - - for (u32 type_index = 0; type_index < props.memoryTypeCount; ++type_index) { - const auto& memory_type{props.memoryTypes[type_index]}; - - if ((memory_type.propertyFlags & flags) != flags) { - // Memory must be device local and host visible - continue; - } - - const auto& heap{props.memoryHeaps[memory_type.heapIndex]}; - if (heap.size >= 7168_MiB) { - // This is the right type of memory - return true; - } - } - - return false; +bool IsStreamHeap(VkMemoryHeap heap) noexcept { + return STREAM_BUFFER_SIZE < (heap.size * 2) / 3; } std::optional<u32> FindMemoryTypeIndex(const VkPhysicalDeviceMemoryProperties& props, u32 type_mask, - VkMemoryPropertyFlags flags, - size_t staging_buffer_size) noexcept { + VkMemoryPropertyFlags flags) noexcept { for (u32 type_index = 0; type_index < props.memoryTypeCount; ++type_index) { if (((type_mask >> type_index) & 1) == 0) { // Memory type is incompatible @@ -69,7 +50,7 @@ std::optional<u32> FindMemoryTypeIndex(const VkPhysicalDeviceMemoryProperties& p // Memory type doesn't have the flags we want continue; } - if (!IsStreamHeap(props.memoryHeaps[memory_type.heapIndex], staging_buffer_size)) { + if (!IsStreamHeap(props.memoryHeaps[memory_type.heapIndex])) { // Memory heap is not suitable for streaming continue; } @@ -80,17 +61,17 @@ std::optional<u32> FindMemoryTypeIndex(const VkPhysicalDeviceMemoryProperties& p } u32 FindMemoryTypeIndex(const VkPhysicalDeviceMemoryProperties& props, u32 type_mask, - bool try_device_local, size_t staging_buffer_size) { + bool try_device_local) { std::optional<u32> type; if (try_device_local) { // Try to find a DEVICE_LOCAL_BIT type, Nvidia and AMD have a dedicated heap for this - type = FindMemoryTypeIndex(props, type_mask, STREAM_FLAGS, staging_buffer_size); + type = FindMemoryTypeIndex(props, type_mask, STREAM_FLAGS); if (type) { return *type; } } // Otherwise try without the DEVICE_LOCAL_BIT - type = FindMemoryTypeIndex(props, type_mask, HOST_FLAGS, staging_buffer_size); + type = FindMemoryTypeIndex(props, type_mask, HOST_FLAGS); if (type) { return *type; } @@ -98,32 +79,20 @@ u32 FindMemoryTypeIndex(const VkPhysicalDeviceMemoryProperties& props, u32 type_ throw vk::Exception(VK_ERROR_OUT_OF_DEVICE_MEMORY); } -size_t Region(size_t iterator, size_t region_size) noexcept { - return iterator / region_size; +size_t Region(size_t iterator) noexcept { + return iterator / REGION_SIZE; } } // Anonymous namespace StagingBufferPool::StagingBufferPool(const Device& device_, MemoryAllocator& memory_allocator_, Scheduler& scheduler_) : device{device_}, memory_allocator{memory_allocator_}, scheduler{scheduler_} { - - const auto memory_properties{device.GetPhysical().GetMemoryProperties().memoryProperties}; - if (HasLargeDeviceLocalHostVisibleMemory(memory_properties)) { - // Possible on many integrated and newer discrete cards - staging_buffer_size = 1_GiB; - } else { - // Well-supported default size used by most Vulkan PC games - staging_buffer_size = 256_MiB; - } - - region_size = staging_buffer_size / StagingBufferPool::NUM_SYNCS; - const vk::Device& dev = device.GetLogical(); stream_buffer = dev.CreateBuffer(VkBufferCreateInfo{ .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, .pNext = nullptr, .flags = 0, - .size = staging_buffer_size, + .size = STREAM_BUFFER_SIZE, .usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT | VK_BUFFER_USAGE_INDEX_BUFFER_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT, .sharingMode = VK_SHARING_MODE_EXCLUSIVE, @@ -148,18 +117,19 @@ StagingBufferPool::StagingBufferPool(const Device& device_, MemoryAllocator& mem .image = nullptr, .buffer = *stream_buffer, }; + const auto memory_properties = device.GetPhysical().GetMemoryProperties().memoryProperties; VkMemoryAllocateInfo stream_memory_info{ .sType = VK_STRUCTURE_TYPE_MEMORY_ALLOCATE_INFO, .pNext = make_dedicated ? &dedicated_info : nullptr, .allocationSize = requirements.size, - .memoryTypeIndex = FindMemoryTypeIndex(memory_properties, requirements.memoryTypeBits, true, - staging_buffer_size), + .memoryTypeIndex = + FindMemoryTypeIndex(memory_properties, requirements.memoryTypeBits, true), }; stream_memory = dev.TryAllocateMemory(stream_memory_info); if (!stream_memory) { LOG_INFO(Render_Vulkan, "Dynamic memory allocation failed, trying with system memory"); - stream_memory_info.memoryTypeIndex = FindMemoryTypeIndex( - memory_properties, requirements.memoryTypeBits, false, staging_buffer_size); + stream_memory_info.memoryTypeIndex = + FindMemoryTypeIndex(memory_properties, requirements.memoryTypeBits, false); stream_memory = dev.AllocateMemory(stream_memory_info); } @@ -167,7 +137,7 @@ StagingBufferPool::StagingBufferPool(const Device& device_, MemoryAllocator& mem stream_memory.SetObjectNameEXT("Stream Buffer Memory"); } stream_buffer.BindMemory(*stream_memory, 0); - stream_pointer = stream_memory.Map(0, staging_buffer_size); + stream_pointer = stream_memory.Map(0, STREAM_BUFFER_SIZE); } StagingBufferPool::~StagingBufferPool() = default; @@ -188,25 +158,25 @@ void StagingBufferPool::TickFrame() { } StagingBufferRef StagingBufferPool::GetStreamBuffer(size_t size) { - if (AreRegionsActive(Region(free_iterator, region_size) + 1, - std::min(Region(iterator + size, region_size) + 1, NUM_SYNCS))) { + if (AreRegionsActive(Region(free_iterator) + 1, + std::min(Region(iterator + size) + 1, NUM_SYNCS))) { // Avoid waiting for the previous usages to be free return GetStagingBuffer(size, MemoryUsage::Upload); } const u64 current_tick = scheduler.CurrentTick(); - std::fill(sync_ticks.begin() + Region(used_iterator, region_size), - sync_ticks.begin() + Region(iterator, region_size), current_tick); + std::fill(sync_ticks.begin() + Region(used_iterator), sync_ticks.begin() + Region(iterator), + current_tick); used_iterator = iterator; free_iterator = std::max(free_iterator, iterator + size); - if (iterator + size >= staging_buffer_size) { - std::fill(sync_ticks.begin() + Region(used_iterator, region_size), - sync_ticks.begin() + NUM_SYNCS, current_tick); + if (iterator + size >= STREAM_BUFFER_SIZE) { + std::fill(sync_ticks.begin() + Region(used_iterator), sync_ticks.begin() + NUM_SYNCS, + current_tick); used_iterator = 0; iterator = 0; free_iterator = size; - if (AreRegionsActive(0, Region(size, region_size) + 1)) { + if (AreRegionsActive(0, Region(size) + 1)) { // Avoid waiting for the previous usages to be free return GetStagingBuffer(size, MemoryUsage::Upload); } diff --git a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h index 90c67177f..91dc84da8 100644 --- a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h +++ b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h @@ -93,9 +93,6 @@ private: size_t free_iterator = 0; std::array<u64, NUM_SYNCS> sync_ticks{}; - size_t staging_buffer_size = 0; - size_t region_size = 0; - StagingBuffersCache device_local_cache; StagingBuffersCache upload_cache; StagingBuffersCache download_cache; diff --git a/src/video_core/renderer_vulkan/vk_state_tracker.cpp b/src/video_core/renderer_vulkan/vk_state_tracker.cpp index f234e1a31..b87c3be66 100644 --- a/src/video_core/renderer_vulkan/vk_state_tracker.cpp +++ b/src/video_core/renderer_vulkan/vk_state_tracker.cpp @@ -51,8 +51,8 @@ Flags MakeInvalidationFlags() { void SetupDirtyViewports(Tables& tables) { FillBlock(tables[0], OFF(viewport_transform), NUM(viewport_transform), Viewports); FillBlock(tables[0], OFF(viewports), NUM(viewports), Viewports); - tables[0][OFF(viewport_transform_enabled)] = Viewports; - tables[1][OFF(screen_y_control)] = Viewports; + tables[0][OFF(viewport_scale_offset_enbled)] = Viewports; + tables[1][OFF(window_origin)] = Viewports; } void SetupDirtyScissors(Tables& tables) { @@ -61,9 +61,9 @@ void SetupDirtyScissors(Tables& tables) { void SetupDirtyDepthBias(Tables& tables) { auto& table = tables[0]; - table[OFF(polygon_offset_units)] = DepthBias; - table[OFF(polygon_offset_clamp)] = DepthBias; - table[OFF(polygon_offset_factor)] = DepthBias; + table[OFF(depth_bias)] = DepthBias; + table[OFF(depth_bias_clamp)] = DepthBias; + table[OFF(slope_scale_depth_bias)] = DepthBias; } void SetupDirtyBlendConstants(Tables& tables) { @@ -77,10 +77,10 @@ void SetupDirtyDepthBounds(Tables& tables) { void SetupDirtyStencilProperties(Tables& tables) { auto& table = tables[0]; table[OFF(stencil_two_side_enable)] = StencilProperties; - table[OFF(stencil_front_func_ref)] = StencilProperties; + table[OFF(stencil_front_ref)] = StencilProperties; table[OFF(stencil_front_mask)] = StencilProperties; table[OFF(stencil_front_func_mask)] = StencilProperties; - table[OFF(stencil_back_func_ref)] = StencilProperties; + table[OFF(stencil_back_ref)] = StencilProperties; table[OFF(stencil_back_mask)] = StencilProperties; table[OFF(stencil_back_func_mask)] = StencilProperties; } @@ -92,8 +92,8 @@ void SetupDirtyLineWidth(Tables& tables) { void SetupDirtyCullMode(Tables& tables) { auto& table = tables[0]; - table[OFF(cull_face)] = CullMode; - table[OFF(cull_test_enabled)] = CullMode; + table[OFF(gl_cull_face)] = CullMode; + table[OFF(gl_cull_test_enabled)] = CullMode; } void SetupDirtyDepthBoundsEnable(Tables& tables) { @@ -114,20 +114,20 @@ void SetupDirtyDepthCompareOp(Tables& tables) { void SetupDirtyFrontFace(Tables& tables) { auto& table = tables[0]; - table[OFF(front_face)] = FrontFace; - table[OFF(screen_y_control)] = FrontFace; + table[OFF(gl_front_face)] = FrontFace; + table[OFF(window_origin)] = FrontFace; } void SetupDirtyStencilOp(Tables& tables) { auto& table = tables[0]; - table[OFF(stencil_front_op_fail)] = StencilOp; - table[OFF(stencil_front_op_zfail)] = StencilOp; - table[OFF(stencil_front_op_zpass)] = StencilOp; - table[OFF(stencil_front_func_func)] = StencilOp; - table[OFF(stencil_back_op_fail)] = StencilOp; - table[OFF(stencil_back_op_zfail)] = StencilOp; - table[OFF(stencil_back_op_zpass)] = StencilOp; - table[OFF(stencil_back_func_func)] = StencilOp; + table[OFF(stencil_front_op.fail)] = StencilOp; + table[OFF(stencil_front_op.zfail)] = StencilOp; + table[OFF(stencil_front_op.zpass)] = StencilOp; + table[OFF(stencil_front_op.func)] = StencilOp; + table[OFF(stencil_back_op.fail)] = StencilOp; + table[OFF(stencil_back_op.zfail)] = StencilOp; + table[OFF(stencil_back_op.zpass)] = StencilOp; + table[OFF(stencil_back_op.func)] = StencilOp; // Table 0 is used by StencilProperties tables[1][OFF(stencil_two_side_enable)] = StencilOp; @@ -139,10 +139,10 @@ void SetupDirtyStencilTestEnable(Tables& tables) { void SetupDirtyBlending(Tables& tables) { tables[0][OFF(color_mask_common)] = Blending; - tables[0][OFF(independent_blend_enable)] = Blending; + tables[0][OFF(blend_per_target_enabled)] = Blending; FillBlock(tables[0], OFF(color_mask), NUM(color_mask), Blending); FillBlock(tables[0], OFF(blend), NUM(blend), Blending); - FillBlock(tables[0], OFF(independent_blend), NUM(independent_blend), Blending); + FillBlock(tables[0], OFF(blend_per_target), NUM(blend_per_target), Blending); } void SetupDirtyViewportSwizzles(Tables& tables) { @@ -166,10 +166,10 @@ void SetupDirtyVertexBindings(Tables& tables) { static constexpr size_t divisor_offset = 3; for (size_t i = 0; i < Regs::NumVertexArrays; ++i) { const u8 flag = static_cast<u8>(VertexBinding0 + i); - tables[0][OFF(instanced_arrays) + i] = VertexInput; - tables[1][OFF(instanced_arrays) + i] = flag; - tables[0][OFF(vertex_array) + i * NUM(vertex_array[0]) + divisor_offset] = VertexInput; - tables[1][OFF(vertex_array) + i * NUM(vertex_array[0]) + divisor_offset] = flag; + tables[0][OFF(vertex_stream_instances) + i] = VertexInput; + tables[1][OFF(vertex_stream_instances) + i] = flag; + tables[0][OFF(vertex_streams) + i * NUM(vertex_streams[0]) + divisor_offset] = VertexInput; + tables[1][OFF(vertex_streams) + i * NUM(vertex_streams[0]) + divisor_offset] = flag; } } } // Anonymous namespace diff --git a/src/video_core/shader_cache.cpp b/src/video_core/shader_cache.cpp index f53066579..d9482371b 100644 --- a/src/video_core/shader_cache.cpp +++ b/src/video_core/shader_cache.cpp @@ -43,14 +43,14 @@ bool ShaderCache::RefreshStages(std::array<u64, 6>& unique_hashes) { } dirty[VideoCommon::Dirty::Shaders] = false; - const GPUVAddr base_addr{maxwell3d->regs.code_address.CodeAddress()}; + const GPUVAddr base_addr{maxwell3d->regs.program_region.Address()}; for (size_t index = 0; index < Tegra::Engines::Maxwell3D::Regs::MaxShaderProgram; ++index) { if (!maxwell3d->regs.IsShaderConfigEnabled(index)) { unique_hashes[index] = 0; continue; } - const auto& shader_config{maxwell3d->regs.shader_config[index]}; - const auto program{static_cast<Tegra::Engines::Maxwell3D::Regs::ShaderProgram>(index)}; + const auto& shader_config{maxwell3d->regs.pipelines[index]}; + const auto program{static_cast<Tegra::Engines::Maxwell3D::Regs::ShaderType>(index)}; const GPUVAddr shader_addr{base_addr + shader_config.offset}; const std::optional<VAddr> cpu_shader_addr{gpu_memory->GpuToCpuAddress(shader_addr)}; if (!cpu_shader_addr) { @@ -90,14 +90,14 @@ const ShaderInfo* ShaderCache::ComputeShader() { void ShaderCache::GetGraphicsEnvironments(GraphicsEnvironments& result, const std::array<u64, NUM_PROGRAMS>& unique_hashes) { size_t env_index{}; - const GPUVAddr base_addr{maxwell3d->regs.code_address.CodeAddress()}; + const GPUVAddr base_addr{maxwell3d->regs.program_region.Address()}; for (size_t index = 0; index < NUM_PROGRAMS; ++index) { if (unique_hashes[index] == 0) { continue; } - const auto program{static_cast<Tegra::Engines::Maxwell3D::Regs::ShaderProgram>(index)}; + const auto program{static_cast<Tegra::Engines::Maxwell3D::Regs::ShaderType>(index)}; auto& env{result.envs[index]}; - const u32 start_address{maxwell3d->regs.shader_config[index].offset}; + const u32 start_address{maxwell3d->regs.pipelines[index].offset}; env = GraphicsEnvironment{*maxwell3d, *gpu_memory, program, base_addr, start_address}; env.SetCachedSize(shader_infos[index]->size_bytes); result.env_ptrs[env_index++] = &env; diff --git a/src/video_core/shader_environment.cpp b/src/video_core/shader_environment.cpp index 5f7625947..fbabb3219 100644 --- a/src/video_core/shader_environment.cpp +++ b/src/video_core/shader_environment.cpp @@ -250,34 +250,34 @@ Shader::TextureType GenericEnvironment::ReadTextureTypeImpl(GPUVAddr tic_addr, u GraphicsEnvironment::GraphicsEnvironment(Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::MemoryManager& gpu_memory_, - Maxwell::ShaderProgram program, GPUVAddr program_base_, + Maxwell::ShaderType program, GPUVAddr program_base_, u32 start_address_) : GenericEnvironment{gpu_memory_, program_base_, start_address_}, maxwell3d{&maxwell3d_} { gpu_memory->ReadBlock(program_base + start_address, &sph, sizeof(sph)); initial_offset = sizeof(sph); - gp_passthrough_mask = maxwell3d->regs.gp_passthrough_mask; + gp_passthrough_mask = maxwell3d->regs.post_vtg_shader_attrib_skip_mask; switch (program) { - case Maxwell::ShaderProgram::VertexA: + case Maxwell::ShaderType::VertexA: stage = Shader::Stage::VertexA; stage_index = 0; break; - case Maxwell::ShaderProgram::VertexB: + case Maxwell::ShaderType::VertexB: stage = Shader::Stage::VertexB; stage_index = 0; break; - case Maxwell::ShaderProgram::TesselationControl: + case Maxwell::ShaderType::TessellationInit: stage = Shader::Stage::TessellationControl; stage_index = 1; break; - case Maxwell::ShaderProgram::TesselationEval: + case Maxwell::ShaderType::Tessellation: stage = Shader::Stage::TessellationEval; stage_index = 2; break; - case Maxwell::ShaderProgram::Geometry: + case Maxwell::ShaderType::Geometry: stage = Shader::Stage::Geometry; stage_index = 3; break; - case Maxwell::ShaderProgram::Fragment: + case Maxwell::ShaderType::Pixel: stage = Shader::Stage::Fragment; stage_index = 4; break; @@ -288,7 +288,7 @@ GraphicsEnvironment::GraphicsEnvironment(Tegra::Engines::Maxwell3D& maxwell3d_, const u64 local_size{sph.LocalMemorySize()}; ASSERT(local_size <= std::numeric_limits<u32>::max()); local_memory_size = static_cast<u32>(local_size) + sph.common3.shader_local_memory_crs_size; - texture_bound = maxwell3d->regs.tex_cb_index; + texture_bound = maxwell3d->regs.bindless_texture_const_buffer_slot; } u32 GraphicsEnvironment::ReadCbufValue(u32 cbuf_index, u32 cbuf_offset) { @@ -304,8 +304,9 @@ u32 GraphicsEnvironment::ReadCbufValue(u32 cbuf_index, u32 cbuf_offset) { Shader::TextureType GraphicsEnvironment::ReadTextureType(u32 handle) { const auto& regs{maxwell3d->regs}; - const bool via_header_index{regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex}; - return ReadTextureTypeImpl(regs.tic.Address(), regs.tic.limit, via_header_index, handle); + const bool via_header_index{regs.sampler_binding == Maxwell::SamplerBinding::ViaHeaderBinding}; + return ReadTextureTypeImpl(regs.tex_header.Address(), regs.tex_header.limit, via_header_index, + handle); } ComputeEnvironment::ComputeEnvironment(Tegra::Engines::KeplerCompute& kepler_compute_, diff --git a/src/video_core/shader_environment.h b/src/video_core/shader_environment.h index 5a145f33a..8b3b8e9f5 100644 --- a/src/video_core/shader_environment.h +++ b/src/video_core/shader_environment.h @@ -93,7 +93,7 @@ public: explicit GraphicsEnvironment() = default; explicit GraphicsEnvironment(Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::MemoryManager& gpu_memory_, - Tegra::Engines::Maxwell3D::Regs::ShaderProgram program, + Tegra::Engines::Maxwell3D::Regs::ShaderType program, GPUVAddr program_base_, u32 start_address_); ~GraphicsEnvironment() override = default; diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp index a2bf08294..6bd133d10 100644 --- a/src/video_core/surface.cpp +++ b/src/video_core/surface.cpp @@ -73,17 +73,17 @@ bool SurfaceTargetIsArray(SurfaceTarget target) { PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) { switch (format) { - case Tegra::DepthFormat::S8_UINT_Z24_UNORM: + case Tegra::DepthFormat::Z24_UNORM_S8_UINT: return PixelFormat::S8_UINT_D24_UNORM; - case Tegra::DepthFormat::D24S8_UNORM: + case Tegra::DepthFormat::S8Z24_UNORM: return PixelFormat::D24_UNORM_S8_UINT; - case Tegra::DepthFormat::D32_FLOAT: + case Tegra::DepthFormat::Z32_FLOAT: return PixelFormat::D32_FLOAT; - case Tegra::DepthFormat::D16_UNORM: + case Tegra::DepthFormat::Z16_UNORM: return PixelFormat::D16_UNORM; case Tegra::DepthFormat::S8_UINT: return PixelFormat::S8_UINT; - case Tegra::DepthFormat::D32_FLOAT_S8X24_UINT: + case Tegra::DepthFormat::Z32_FLOAT_X24S8_UINT: return PixelFormat::D32_FLOAT_S8_UINT; default: UNIMPLEMENTED_MSG("Unimplemented format={}", format); @@ -117,9 +117,9 @@ PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) return PixelFormat::R32G32_UINT; case Tegra::RenderTargetFormat::R16G16B16X16_FLOAT: return PixelFormat::R16G16B16X16_FLOAT; - case Tegra::RenderTargetFormat::B8G8R8A8_UNORM: + case Tegra::RenderTargetFormat::A8R8G8B8_UNORM: return PixelFormat::B8G8R8A8_UNORM; - case Tegra::RenderTargetFormat::B8G8R8A8_SRGB: + case Tegra::RenderTargetFormat::A8R8G8B8_SRGB: return PixelFormat::B8G8R8A8_SRGB; case Tegra::RenderTargetFormat::A2B10G10R10_UNORM: return PixelFormat::A2B10G10R10_UNORM; diff --git a/src/video_core/texture_cache/image_info.cpp b/src/video_core/texture_cache/image_info.cpp index 6c073ee57..852ec2519 100644 --- a/src/video_core/texture_cache/image_info.cpp +++ b/src/video_core/texture_cache/image_info.cpp @@ -1,6 +1,8 @@ // SPDX-FileCopyrightText: Copyright 2020 yuzu Emulator Project // SPDX-License-Identifier: GPL-2.0-or-later +#include <fmt/format.h> + #include "common/assert.h" #include "video_core/surface.h" #include "video_core/texture_cache/format_lookup_table.h" @@ -12,6 +14,7 @@ namespace VideoCommon { +using Tegra::Engines::Maxwell3D; using Tegra::Texture::TextureType; using Tegra::Texture::TICEntry; using VideoCore::Surface::PixelFormat; @@ -107,12 +110,13 @@ ImageInfo::ImageInfo(const TICEntry& config) noexcept { } } -ImageInfo::ImageInfo(const Tegra::Engines::Maxwell3D::Regs& regs, size_t index) noexcept { +ImageInfo::ImageInfo(const Maxwell3D::Regs& regs, size_t index) noexcept { const auto& rt = regs.rt[index]; format = VideoCore::Surface::PixelFormatFromRenderTargetFormat(rt.format); rescaleable = false; if (rt.tile_mode.is_pitch_linear) { - ASSERT(rt.tile_mode.is_3d == 0); + ASSERT(rt.tile_mode.dim_control == + Maxwell3D::Regs::TileMode::DimensionControl::DepthDefinesArray); type = ImageType::Linear; pitch = rt.width; size = Extent3D{ @@ -124,15 +128,16 @@ ImageInfo::ImageInfo(const Tegra::Engines::Maxwell3D::Regs& regs, size_t index) } size.width = rt.width; size.height = rt.height; - layer_stride = rt.layer_stride * 4; + layer_stride = rt.array_pitch * 4; maybe_unaligned_layer_stride = layer_stride; - num_samples = NumSamples(regs.multisample_mode); + num_samples = NumSamples(regs.anti_alias_samples_mode); block = Extent3D{ .width = rt.tile_mode.block_width, .height = rt.tile_mode.block_height, .depth = rt.tile_mode.block_depth, }; - if (rt.tile_mode.is_3d) { + if (rt.tile_mode.dim_control == + Maxwell3D::Regs::TileMode::DimensionControl::DepthDefinesDepth) { type = ImageType::e3D; size.depth = rt.depth; } else { @@ -146,31 +151,37 @@ ImageInfo::ImageInfo(const Tegra::Engines::Maxwell3D::Regs& regs, size_t index) ImageInfo::ImageInfo(const Tegra::Engines::Maxwell3D::Regs& regs) noexcept { format = VideoCore::Surface::PixelFormatFromDepthFormat(regs.zeta.format); - size.width = regs.zeta_width; - size.height = regs.zeta_height; + size.width = regs.zeta_size.width; + size.height = regs.zeta_size.height; rescaleable = false; resources.levels = 1; - layer_stride = regs.zeta.layer_stride * 4; + layer_stride = regs.zeta.array_pitch * 4; maybe_unaligned_layer_stride = layer_stride; - num_samples = NumSamples(regs.multisample_mode); + num_samples = NumSamples(regs.anti_alias_samples_mode); block = Extent3D{ .width = regs.zeta.tile_mode.block_width, .height = regs.zeta.tile_mode.block_height, .depth = regs.zeta.tile_mode.block_depth, }; if (regs.zeta.tile_mode.is_pitch_linear) { - ASSERT(regs.zeta.tile_mode.is_3d == 0); + ASSERT(regs.zeta.tile_mode.dim_control == + Maxwell3D::Regs::TileMode::DimensionControl::DepthDefinesArray); type = ImageType::Linear; pitch = size.width * BytesPerBlock(format); - } else if (regs.zeta.tile_mode.is_3d) { + } else if (regs.zeta.tile_mode.dim_control == + Maxwell3D::Regs::TileMode::DimensionControl::DepthDefinesDepth) { ASSERT(regs.zeta.tile_mode.is_pitch_linear == 0); + ASSERT(regs.zeta_size.dim_control == + Maxwell3D::Regs::ZetaSize::DimensionControl::ArraySizeOne); type = ImageType::e3D; - size.depth = regs.zeta_depth; + size.depth = regs.zeta_size.depth; } else { + ASSERT(regs.zeta_size.dim_control == + Maxwell3D::Regs::ZetaSize::DimensionControl::DepthDefinesArray); rescaleable = block.depth == 0; downscaleable = size.height > 512; type = ImageType::e2D; - resources.layers = regs.zeta_depth; + resources.layers = regs.zeta_size.depth; } } diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index eaf4a1c95..413baf730 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -189,15 +189,16 @@ typename P::Sampler* TextureCache<P>::GetComputeSampler(u32 index) { template <class P> void TextureCache<P>::SynchronizeGraphicsDescriptors() { - using SamplerIndex = Tegra::Engines::Maxwell3D::Regs::SamplerIndex; - const bool linked_tsc = maxwell3d->regs.sampler_index == SamplerIndex::ViaHeaderIndex; - const u32 tic_limit = maxwell3d->regs.tic.limit; - const u32 tsc_limit = linked_tsc ? tic_limit : maxwell3d->regs.tsc.limit; - if (channel_state->graphics_sampler_table.Synchornize(maxwell3d->regs.tsc.Address(), + using SamplerBinding = Tegra::Engines::Maxwell3D::Regs::SamplerBinding; + const bool linked_tsc = maxwell3d->regs.sampler_binding == SamplerBinding::ViaHeaderBinding; + const u32 tic_limit = maxwell3d->regs.tex_header.limit; + const u32 tsc_limit = linked_tsc ? tic_limit : maxwell3d->regs.tex_sampler.limit; + if (channel_state->graphics_sampler_table.Synchornize(maxwell3d->regs.tex_sampler.Address(), tsc_limit)) { channel_state->graphics_sampler_ids.resize(tsc_limit + 1, CORRUPT_ID); } - if (channel_state->graphics_image_table.Synchornize(maxwell3d->regs.tic.Address(), tic_limit)) { + if (channel_state->graphics_image_table.Synchornize(maxwell3d->regs.tex_header.Address(), + tic_limit)) { channel_state->graphics_image_view_ids.resize(tic_limit + 1, CORRUPT_ID); } } @@ -352,8 +353,8 @@ void TextureCache<P>::UpdateRenderTargets(bool is_clear) { down_shift = Settings::values.resolution_info.down_shift; } render_targets.size = Extent2D{ - (maxwell3d->regs.render_area.width * up_scale) >> down_shift, - (maxwell3d->regs.render_area.height * up_scale) >> down_shift, + (maxwell3d->regs.surface_clip.width * up_scale) >> down_shift, + (maxwell3d->regs.surface_clip.height * up_scale) >> down_shift, }; render_targets.is_rescaled = is_rescaling; @@ -1980,7 +1981,7 @@ bool TextureCache<P>::IsFullClear(ImageViewId id) { // Images with multiple resources can't be cleared in a single call return false; } - if (regs.clear_flags.scissor == 0) { + if (regs.clear_control.use_scissor == 0) { // If scissor testing is disabled, the clear is always full return true; } diff --git a/src/video_core/transform_feedback.cpp b/src/video_core/transform_feedback.cpp index 7e605981c..45071185a 100644 --- a/src/video_core/transform_feedback.cpp +++ b/src/video_core/transform_feedback.cpp @@ -15,51 +15,51 @@ namespace VideoCommon { std::vector<Shader::TransformFeedbackVarying> MakeTransformFeedbackVaryings( const TransformFeedbackState& state) { static constexpr std::array VECTORS{ - 28, // gl_Position - 32, // Generic 0 - 36, // Generic 1 - 40, // Generic 2 - 44, // Generic 3 - 48, // Generic 4 - 52, // Generic 5 - 56, // Generic 6 - 60, // Generic 7 - 64, // Generic 8 - 68, // Generic 9 - 72, // Generic 10 - 76, // Generic 11 - 80, // Generic 12 - 84, // Generic 13 - 88, // Generic 14 - 92, // Generic 15 - 96, // Generic 16 - 100, // Generic 17 - 104, // Generic 18 - 108, // Generic 19 - 112, // Generic 20 - 116, // Generic 21 - 120, // Generic 22 - 124, // Generic 23 - 128, // Generic 24 - 132, // Generic 25 - 136, // Generic 26 - 140, // Generic 27 - 144, // Generic 28 - 148, // Generic 29 - 152, // Generic 30 - 156, // Generic 31 - 160, // gl_FrontColor - 164, // gl_FrontSecondaryColor - 160, // gl_BackColor - 164, // gl_BackSecondaryColor - 192, // gl_TexCoord[0] - 196, // gl_TexCoord[1] - 200, // gl_TexCoord[2] - 204, // gl_TexCoord[3] - 208, // gl_TexCoord[4] - 212, // gl_TexCoord[5] - 216, // gl_TexCoord[6] - 220, // gl_TexCoord[7] + 28U, // gl_Position + 32U, // Generic 0 + 36U, // Generic 1 + 40U, // Generic 2 + 44U, // Generic 3 + 48U, // Generic 4 + 52U, // Generic 5 + 56U, // Generic 6 + 60U, // Generic 7 + 64U, // Generic 8 + 68U, // Generic 9 + 72U, // Generic 10 + 76U, // Generic 11 + 80U, // Generic 12 + 84U, // Generic 13 + 88U, // Generic 14 + 92U, // Generic 15 + 96U, // Generic 16 + 100U, // Generic 17 + 104U, // Generic 18 + 108U, // Generic 19 + 112U, // Generic 20 + 116U, // Generic 21 + 120U, // Generic 22 + 124U, // Generic 23 + 128U, // Generic 24 + 132U, // Generic 25 + 136U, // Generic 26 + 140U, // Generic 27 + 144U, // Generic 28 + 148U, // Generic 29 + 152U, // Generic 30 + 156U, // Generic 31 + 160U, // gl_FrontColor + 164U, // gl_FrontSecondaryColor + 160U, // gl_BackColor + 164U, // gl_BackSecondaryColor + 192U, // gl_TexCoord[0] + 196U, // gl_TexCoord[1] + 200U, // gl_TexCoord[2] + 204U, // gl_TexCoord[3] + 208U, // gl_TexCoord[4] + 212U, // gl_TexCoord[5] + 216U, // gl_TexCoord[6] + 220U, // gl_TexCoord[7] }; std::vector<Shader::TransformFeedbackVarying> xfb(256); for (size_t buffer = 0; buffer < state.layouts.size(); ++buffer) { @@ -68,8 +68,20 @@ std::vector<Shader::TransformFeedbackVarying> MakeTransformFeedbackVaryings( const u32 varying_count = layout.varying_count; u32 highest = 0; for (u32 offset = 0; offset < varying_count; ++offset) { - const u32 base_offset = offset; - const u8 location = locations[offset]; + const auto get_attribute = [&locations](u32 index) -> u32 { + switch (index % 4) { + case 0: + return locations[index / 4].attribute0.Value(); + case 1: + return locations[index / 4].attribute1.Value(); + case 2: + return locations[index / 4].attribute2.Value(); + case 3: + return locations[index / 4].attribute3.Value(); + } + UNREACHABLE(); + return 0; + }; UNIMPLEMENTED_IF_MSG(layout.stream != 0, "Stream is not zero: {}", layout.stream); Shader::TransformFeedbackVarying varying{ @@ -78,16 +90,18 @@ std::vector<Shader::TransformFeedbackVarying> MakeTransformFeedbackVaryings( .offset = offset * 4, .components = 1, }; - if (std::ranges::find(VECTORS, Common::AlignDown(location, 4)) != VECTORS.end()) { - UNIMPLEMENTED_IF_MSG(location % 4 != 0, "Unaligned TFB"); + const u32 base_offset = offset; + const auto attribute{get_attribute(offset)}; + if (std::ranges::find(VECTORS, Common::AlignDown(attribute, 4)) != VECTORS.end()) { + UNIMPLEMENTED_IF_MSG(attribute % 4 != 0, "Unaligned TFB {}", attribute); - const u8 base_index = location / 4; - while (offset + 1 < varying_count && base_index == locations[offset + 1] / 4) { + const auto base_index = attribute / 4; + while (offset + 1 < varying_count && base_index == get_attribute(offset + 1) / 4) { ++offset; ++varying.components; } } - xfb[location] = varying; + xfb[attribute] = varying; highest = std::max(highest, (base_offset + varying.components) * 4); } UNIMPLEMENTED_IF(highest != layout.stride); diff --git a/src/video_core/transform_feedback.h b/src/video_core/transform_feedback.h index a519adb59..d13eb16c3 100644 --- a/src/video_core/transform_feedback.h +++ b/src/video_core/transform_feedback.h @@ -19,7 +19,8 @@ struct TransformFeedbackState { u32 stride; }; std::array<Layout, Tegra::Engines::Maxwell3D::Regs::NumTransformFeedbackBuffers> layouts; - std::array<std::array<u8, 128>, Tegra::Engines::Maxwell3D::Regs::NumTransformFeedbackBuffers> + std::array<std::array<Tegra::Engines::Maxwell3D::Regs::StreamOutLayout, 32>, + Tegra::Engines::Maxwell3D::Regs::NumTransformFeedbackBuffers> varyings; }; diff --git a/src/yuzu/configuration/configure_graphics.cpp b/src/yuzu/configuration/configure_graphics.cpp index 87e5d0f48..bd69d04a6 100644 --- a/src/yuzu/configuration/configure_graphics.cpp +++ b/src/yuzu/configuration/configure_graphics.cpp @@ -57,9 +57,10 @@ ConfigureGraphics::ConfigureGraphics(const Core::System& system_, QWidget* paren UpdateBackgroundColorButton(new_bg_color); }); - ui->api->setEnabled(!UISettings::values.has_broken_vulkan); - ui->api_widget->setEnabled(!UISettings::values.has_broken_vulkan || - Settings::IsConfiguringGlobal()); + ui->api->setEnabled(!UISettings::values.has_broken_vulkan && ui->api->isEnabled()); + ui->api_widget->setEnabled( + (!UISettings::values.has_broken_vulkan || Settings::IsConfiguringGlobal()) && + ui->api_widget->isEnabled()); ui->bg_label->setVisible(Settings::IsConfiguringGlobal()); ui->bg_combobox->setVisible(!Settings::IsConfiguringGlobal()); } diff --git a/src/yuzu/configuration/configure_graphics.ui b/src/yuzu/configuration/configure_graphics.ui index 1e4f74704..fdbb33372 100644 --- a/src/yuzu/configuration/configure_graphics.ui +++ b/src/yuzu/configuration/configure_graphics.ui @@ -301,6 +301,11 @@ </item> <item> <property name="text"> + <string>Force 16:10</string> + </property> + </item> + <item> + <property name="text"> <string>Stretch to Window</string> </property> </item> diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index 4146ebc2c..f45a25410 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -262,6 +262,18 @@ static QString PrettyProductName() { return QSysInfo::prettyProductName(); } +#ifdef _WIN32 +static void OverrideWindowsFont() { + // Qt5 chooses these fonts on Windows and they have fairly ugly alphanumeric/cyrllic characters + // Asking to use "MS Shell Dlg 2" gives better other chars while leaving the Chinese Characters. + const QString startup_font = QApplication::font().family(); + const QStringList ugly_fonts = {QStringLiteral("SimSun"), QStringLiteral("PMingLiU")}; + if (ugly_fonts.contains(startup_font)) { + QApplication::setFont(QFont(QStringLiteral("MS Shell Dlg 2"), 9, QFont::Normal)); + } +} +#endif + bool GMainWindow::CheckDarkMode() { #ifdef __linux__ const QPalette test_palette(qApp->palette()); @@ -4137,6 +4149,10 @@ int main(int argc, char* argv[]) { QCoreApplication::setAttribute(Qt::AA_DontCheckOpenGLContextThreadAffinity); QApplication app(argc, argv); +#ifdef _WIN32 + OverrideWindowsFont(); +#endif + // Workaround for QTBUG-85409, for Suzhou numerals the number 1 is actually \u3021 // so we can see if we get \u3008 instead // TL;DR all other number formats are consecutive in unicode code points |