summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--src/audio_core/codec.cpp4
-rw-r--r--src/audio_core/cubeb_sink.cpp8
-rw-r--r--src/common/swap.h6
-rw-r--r--src/core/hle/service/nvflinger/nvflinger.cpp76
-rw-r--r--src/core/hle/service/nvflinger/nvflinger.h11
-rw-r--r--src/core/hle/service/vi/display/vi_display.cpp49
-rw-r--r--src/core/hle/service/vi/display/vi_display.h74
-rw-r--r--src/core/hle/service/vi/layer/vi_layer.cpp3
-rw-r--r--src/core/hle/service/vi/layer/vi_layer.h37
-rw-r--r--src/core/hle/service/vi/vi.cpp27
-rw-r--r--src/video_core/engines/shader_bytecode.h6
-rw-r--r--src/video_core/engines/shader_header.h41
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.cpp6
-rw-r--r--src/video_core/renderer_opengl/gl_shader_decompiler.cpp44
-rw-r--r--src/video_core/renderer_opengl/gl_shader_gen.cpp4
-rw-r--r--src/video_core/shader/decode/memory.cpp2
-rw-r--r--src/video_core/shader/decode/other.cpp15
-rw-r--r--src/video_core/shader/track.cpp10
18 files changed, 303 insertions, 120 deletions
diff --git a/src/audio_core/codec.cpp b/src/audio_core/codec.cpp
index 454de798b..c5a0d98ce 100644
--- a/src/audio_core/codec.cpp
+++ b/src/audio_core/codec.cpp
@@ -68,8 +68,8 @@ std::vector<s16> DecodeADPCM(const u8* const data, std::size_t size, const ADPCM
}
}
- state.yn1 = yn1;
- state.yn2 = yn2;
+ state.yn1 = static_cast<s16>(yn1);
+ state.yn2 = static_cast<s16>(yn2);
return ret;
}
diff --git a/src/audio_core/cubeb_sink.cpp b/src/audio_core/cubeb_sink.cpp
index 097328901..dc45dedd3 100644
--- a/src/audio_core/cubeb_sink.cpp
+++ b/src/audio_core/cubeb_sink.cpp
@@ -46,7 +46,7 @@ public:
}
}
- ~CubebSinkStream() {
+ ~CubebSinkStream() override {
if (!ctx) {
return;
}
@@ -75,11 +75,11 @@ public:
queue.Push(samples);
}
- std::size_t SamplesInQueue(u32 num_channels) const override {
+ std::size_t SamplesInQueue(u32 channel_count) const override {
if (!ctx)
return 0;
- return queue.Size() / num_channels;
+ return queue.Size() / channel_count;
}
void Flush() override {
@@ -98,7 +98,7 @@ private:
u32 num_channels{};
Common::RingBuffer<s16, 0x10000> queue;
- std::array<s16, 2> last_frame;
+ std::array<s16, 2> last_frame{};
std::atomic<bool> should_flush{};
TimeStretcher time_stretch;
diff --git a/src/common/swap.h b/src/common/swap.h
index 32af0b6ac..0e219747f 100644
--- a/src/common/swap.h
+++ b/src/common/swap.h
@@ -28,8 +28,8 @@
#include <cstring>
#include "common/common_types.h"
-// GCC 4.6+
-#if __GNUC__ >= 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 6)
+// GCC
+#ifdef __GNUC__
#if __BYTE_ORDER__ && (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) && !defined(COMMON_LITTLE_ENDIAN)
#define COMMON_LITTLE_ENDIAN 1
@@ -38,7 +38,7 @@
#endif
// LLVM/clang
-#elif __clang__
+#elif defined(__clang__)
#if __LITTLE_ENDIAN__ && !defined(COMMON_LITTLE_ENDIAN)
#define COMMON_LITTLE_ENDIAN 1
diff --git a/src/core/hle/service/nvflinger/nvflinger.cpp b/src/core/hle/service/nvflinger/nvflinger.cpp
index b5d452db1..56f31e2ac 100644
--- a/src/core/hle/service/nvflinger/nvflinger.cpp
+++ b/src/core/hle/service/nvflinger/nvflinger.cpp
@@ -28,9 +28,13 @@ namespace Service::NVFlinger {
constexpr std::size_t SCREEN_REFRESH_RATE = 60;
constexpr u64 frame_ticks = static_cast<u64>(Core::Timing::BASE_CLOCK_RATE / SCREEN_REFRESH_RATE);
-NVFlinger::NVFlinger(Core::Timing::CoreTiming& core_timing)
- : displays{{0, "Default"}, {1, "External"}, {2, "Edid"}, {3, "Internal"}, {4, "Null"}},
- core_timing{core_timing} {
+NVFlinger::NVFlinger(Core::Timing::CoreTiming& core_timing) : core_timing{core_timing} {
+ displays.emplace_back(0, "Default");
+ displays.emplace_back(1, "External");
+ displays.emplace_back(2, "Edid");
+ displays.emplace_back(3, "Internal");
+ displays.emplace_back(4, "Null");
+
// Schedule the screen composition events
composition_event =
core_timing.RegisterEvent("ScreenComposition", [this](u64 userdata, int cycles_late) {
@@ -55,13 +59,14 @@ std::optional<u64> NVFlinger::OpenDisplay(std::string_view name) {
// TODO(Subv): Currently we only support the Default display.
ASSERT(name == "Default");
- const auto itr = std::find_if(displays.begin(), displays.end(),
- [&](const VI::Display& display) { return display.name == name; });
+ const auto itr =
+ std::find_if(displays.begin(), displays.end(),
+ [&](const VI::Display& display) { return display.GetName() == name; });
if (itr == displays.end()) {
return {};
}
- return itr->id;
+ return itr->GetID();
}
std::optional<u64> NVFlinger::CreateLayer(u64 display_id) {
@@ -71,13 +76,10 @@ std::optional<u64> NVFlinger::CreateLayer(u64 display_id) {
return {};
}
- ASSERT_MSG(display->layers.empty(), "Only one layer is supported per display at the moment");
-
const u64 layer_id = next_layer_id++;
const u32 buffer_queue_id = next_buffer_queue_id++;
- auto buffer_queue = std::make_shared<BufferQueue>(buffer_queue_id, layer_id);
- display->layers.emplace_back(layer_id, buffer_queue);
- buffer_queues.emplace_back(std::move(buffer_queue));
+ buffer_queues.emplace_back(buffer_queue_id, layer_id);
+ display->CreateLayer(layer_id, buffer_queues.back());
return layer_id;
}
@@ -88,7 +90,7 @@ std::optional<u32> NVFlinger::FindBufferQueueId(u64 display_id, u64 layer_id) co
return {};
}
- return layer->buffer_queue->GetId();
+ return layer->GetBufferQueue().GetId();
}
Kernel::SharedPtr<Kernel::ReadableEvent> NVFlinger::FindVsyncEvent(u64 display_id) const {
@@ -98,12 +100,20 @@ Kernel::SharedPtr<Kernel::ReadableEvent> NVFlinger::FindVsyncEvent(u64 display_i
return nullptr;
}
- return display->vsync_event.readable;
+ return display->GetVSyncEvent();
}
-std::shared_ptr<BufferQueue> NVFlinger::FindBufferQueue(u32 id) const {
+BufferQueue& NVFlinger::FindBufferQueue(u32 id) {
const auto itr = std::find_if(buffer_queues.begin(), buffer_queues.end(),
- [&](const auto& queue) { return queue->GetId() == id; });
+ [id](const auto& queue) { return queue.GetId() == id; });
+
+ ASSERT(itr != buffer_queues.end());
+ return *itr;
+}
+
+const BufferQueue& NVFlinger::FindBufferQueue(u32 id) const {
+ const auto itr = std::find_if(buffer_queues.begin(), buffer_queues.end(),
+ [id](const auto& queue) { return queue.GetId() == id; });
ASSERT(itr != buffer_queues.end());
return *itr;
@@ -112,7 +122,7 @@ std::shared_ptr<BufferQueue> NVFlinger::FindBufferQueue(u32 id) const {
VI::Display* NVFlinger::FindDisplay(u64 display_id) {
const auto itr =
std::find_if(displays.begin(), displays.end(),
- [&](const VI::Display& display) { return display.id == display_id; });
+ [&](const VI::Display& display) { return display.GetID() == display_id; });
if (itr == displays.end()) {
return nullptr;
@@ -124,7 +134,7 @@ VI::Display* NVFlinger::FindDisplay(u64 display_id) {
const VI::Display* NVFlinger::FindDisplay(u64 display_id) const {
const auto itr =
std::find_if(displays.begin(), displays.end(),
- [&](const VI::Display& display) { return display.id == display_id; });
+ [&](const VI::Display& display) { return display.GetID() == display_id; });
if (itr == displays.end()) {
return nullptr;
@@ -140,14 +150,7 @@ VI::Layer* NVFlinger::FindLayer(u64 display_id, u64 layer_id) {
return nullptr;
}
- const auto itr = std::find_if(display->layers.begin(), display->layers.end(),
- [&](const VI::Layer& layer) { return layer.id == layer_id; });
-
- if (itr == display->layers.end()) {
- return nullptr;
- }
-
- return &*itr;
+ return display->FindLayer(layer_id);
}
const VI::Layer* NVFlinger::FindLayer(u64 display_id, u64 layer_id) const {
@@ -157,33 +160,24 @@ const VI::Layer* NVFlinger::FindLayer(u64 display_id, u64 layer_id) const {
return nullptr;
}
- const auto itr = std::find_if(display->layers.begin(), display->layers.end(),
- [&](const VI::Layer& layer) { return layer.id == layer_id; });
-
- if (itr == display->layers.end()) {
- return nullptr;
- }
-
- return &*itr;
+ return display->FindLayer(layer_id);
}
void NVFlinger::Compose() {
for (auto& display : displays) {
// Trigger vsync for this display at the end of drawing
- SCOPE_EXIT({ display.vsync_event.writable->Signal(); });
+ SCOPE_EXIT({ display.SignalVSyncEvent(); });
// Don't do anything for displays without layers.
- if (display.layers.empty())
+ if (!display.HasLayers())
continue;
// TODO(Subv): Support more than 1 layer.
- ASSERT_MSG(display.layers.size() == 1, "Max 1 layer per display is supported");
-
- VI::Layer& layer = display.layers[0];
- auto& buffer_queue = layer.buffer_queue;
+ VI::Layer& layer = display.GetLayer(0);
+ auto& buffer_queue = layer.GetBufferQueue();
// Search for a queued buffer and acquire it
- auto buffer = buffer_queue->AcquireBuffer();
+ auto buffer = buffer_queue.AcquireBuffer();
MicroProfileFlip();
@@ -208,7 +202,7 @@ void NVFlinger::Compose() {
igbp_buffer.width, igbp_buffer.height, igbp_buffer.stride,
buffer->get().transform, buffer->get().crop_rect);
- buffer_queue->ReleaseBuffer(buffer->get().slot);
+ buffer_queue.ReleaseBuffer(buffer->get().slot);
}
}
diff --git a/src/core/hle/service/nvflinger/nvflinger.h b/src/core/hle/service/nvflinger/nvflinger.h
index 2e000af91..c0a83fffb 100644
--- a/src/core/hle/service/nvflinger/nvflinger.h
+++ b/src/core/hle/service/nvflinger/nvflinger.h
@@ -28,8 +28,8 @@ class Module;
} // namespace Service::Nvidia
namespace Service::VI {
-struct Display;
-struct Layer;
+class Display;
+class Layer;
} // namespace Service::VI
namespace Service::NVFlinger {
@@ -65,7 +65,10 @@ public:
Kernel::SharedPtr<Kernel::ReadableEvent> FindVsyncEvent(u64 display_id) const;
/// Obtains a buffer queue identified by the ID.
- std::shared_ptr<BufferQueue> FindBufferQueue(u32 id) const;
+ BufferQueue& FindBufferQueue(u32 id);
+
+ /// Obtains a buffer queue identified by the ID.
+ const BufferQueue& FindBufferQueue(u32 id) const;
/// Performs a composition request to the emulated nvidia GPU and triggers the vsync events when
/// finished.
@@ -87,7 +90,7 @@ private:
std::shared_ptr<Nvidia::Module> nvdrv;
std::vector<VI::Display> displays;
- std::vector<std::shared_ptr<BufferQueue>> buffer_queues;
+ std::vector<BufferQueue> buffer_queues;
/// Id to use for the next layer that is created, this counter is shared among all displays.
u64 next_layer_id = 1;
diff --git a/src/core/hle/service/vi/display/vi_display.cpp b/src/core/hle/service/vi/display/vi_display.cpp
index a108e468f..01d80311b 100644
--- a/src/core/hle/service/vi/display/vi_display.cpp
+++ b/src/core/hle/service/vi/display/vi_display.cpp
@@ -2,8 +2,12 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
+#include <algorithm>
+#include <utility>
+
#include <fmt/format.h>
+#include "common/assert.h"
#include "core/core.h"
#include "core/hle/kernel/readable_event.h"
#include "core/hle/service/vi/display/vi_display.h"
@@ -19,4 +23,49 @@ Display::Display(u64 id, std::string name) : id{id}, name{std::move(name)} {
Display::~Display() = default;
+Layer& Display::GetLayer(std::size_t index) {
+ return layers.at(index);
+}
+
+const Layer& Display::GetLayer(std::size_t index) const {
+ return layers.at(index);
+}
+
+Kernel::SharedPtr<Kernel::ReadableEvent> Display::GetVSyncEvent() const {
+ return vsync_event.readable;
+}
+
+void Display::SignalVSyncEvent() {
+ vsync_event.writable->Signal();
+}
+
+void Display::CreateLayer(u64 id, NVFlinger::BufferQueue& buffer_queue) {
+ // TODO(Subv): Support more than 1 layer.
+ ASSERT_MSG(layers.empty(), "Only one layer is supported per display at the moment");
+
+ layers.emplace_back(id, buffer_queue);
+}
+
+Layer* Display::FindLayer(u64 id) {
+ const auto itr = std::find_if(layers.begin(), layers.end(),
+ [id](const VI::Layer& layer) { return layer.GetID() == id; });
+
+ if (itr == layers.end()) {
+ return nullptr;
+ }
+
+ return &*itr;
+}
+
+const Layer* Display::FindLayer(u64 id) const {
+ const auto itr = std::find_if(layers.begin(), layers.end(),
+ [id](const VI::Layer& layer) { return layer.GetID() == id; });
+
+ if (itr == layers.end()) {
+ return nullptr;
+ }
+
+ return &*itr;
+}
+
} // namespace Service::VI
diff --git a/src/core/hle/service/vi/display/vi_display.h b/src/core/hle/service/vi/display/vi_display.h
index df44db306..2acd46ff8 100644
--- a/src/core/hle/service/vi/display/vi_display.h
+++ b/src/core/hle/service/vi/display/vi_display.h
@@ -10,14 +10,84 @@
#include "common/common_types.h"
#include "core/hle/kernel/writable_event.h"
+namespace Service::NVFlinger {
+class BufferQueue;
+}
+
namespace Service::VI {
-struct Layer;
+class Layer;
-struct Display {
+/// Represents a single display type
+class Display {
+public:
+ /// Constructs a display with a given unique ID and name.
+ ///
+ /// @param id The unique ID for this display.
+ /// @param name The name for this display.
+ ///
Display(u64 id, std::string name);
~Display();
+ Display(const Display&) = delete;
+ Display& operator=(const Display&) = delete;
+
+ Display(Display&&) = default;
+ Display& operator=(Display&&) = default;
+
+ /// Gets the unique ID assigned to this display.
+ u64 GetID() const {
+ return id;
+ }
+
+ /// Gets the name of this display
+ const std::string& GetName() const {
+ return name;
+ }
+
+ /// Whether or not this display has any layers added to it.
+ bool HasLayers() const {
+ return !layers.empty();
+ }
+
+ /// Gets a layer for this display based off an index.
+ Layer& GetLayer(std::size_t index);
+
+ /// Gets a layer for this display based off an index.
+ const Layer& GetLayer(std::size_t index) const;
+
+ /// Gets the readable vsync event.
+ Kernel::SharedPtr<Kernel::ReadableEvent> GetVSyncEvent() const;
+
+ /// Signals the internal vsync event.
+ void SignalVSyncEvent();
+
+ /// Creates and adds a layer to this display with the given ID.
+ ///
+ /// @param id The ID to assign to the created layer.
+ /// @param buffer_queue The buffer queue for the layer instance to use.
+ ///
+ void CreateLayer(u64 id, NVFlinger::BufferQueue& buffer_queue);
+
+ /// Attempts to find a layer with the given ID.
+ ///
+ /// @param id The layer ID.
+ ///
+ /// @returns If found, the Layer instance with the given ID.
+ /// If not found, then nullptr is returned.
+ ///
+ Layer* FindLayer(u64 id);
+
+ /// Attempts to find a layer with the given ID.
+ ///
+ /// @param id The layer ID.
+ ///
+ /// @returns If found, the Layer instance with the given ID.
+ /// If not found, then nullptr is returned.
+ ///
+ const Layer* FindLayer(u64 id) const;
+
+private:
u64 id;
std::string name;
diff --git a/src/core/hle/service/vi/layer/vi_layer.cpp b/src/core/hle/service/vi/layer/vi_layer.cpp
index 3a83e5b95..954225c26 100644
--- a/src/core/hle/service/vi/layer/vi_layer.cpp
+++ b/src/core/hle/service/vi/layer/vi_layer.cpp
@@ -6,8 +6,7 @@
namespace Service::VI {
-Layer::Layer(u64 id, std::shared_ptr<NVFlinger::BufferQueue> queue)
- : id{id}, buffer_queue{std::move(queue)} {}
+Layer::Layer(u64 id, NVFlinger::BufferQueue& queue) : id{id}, buffer_queue{queue} {}
Layer::~Layer() = default;
diff --git a/src/core/hle/service/vi/layer/vi_layer.h b/src/core/hle/service/vi/layer/vi_layer.h
index df328e09f..c6bfd01f6 100644
--- a/src/core/hle/service/vi/layer/vi_layer.h
+++ b/src/core/hle/service/vi/layer/vi_layer.h
@@ -4,8 +4,6 @@
#pragma once
-#include <memory>
-
#include "common/common_types.h"
namespace Service::NVFlinger {
@@ -14,12 +12,41 @@ class BufferQueue;
namespace Service::VI {
-struct Layer {
- Layer(u64 id, std::shared_ptr<NVFlinger::BufferQueue> queue);
+/// Represents a single display layer.
+class Layer {
+public:
+ /// Constructs a layer with a given ID and buffer queue.
+ ///
+ /// @param id The ID to assign to this layer.
+ /// @param queue The buffer queue for this layer to use.
+ ///
+ Layer(u64 id, NVFlinger::BufferQueue& queue);
~Layer();
+ Layer(const Layer&) = delete;
+ Layer& operator=(const Layer&) = delete;
+
+ Layer(Layer&&) = default;
+ Layer& operator=(Layer&&) = delete;
+
+ /// Gets the ID for this layer.
+ u64 GetID() const {
+ return id;
+ }
+
+ /// Gets a reference to the buffer queue this layer is using.
+ NVFlinger::BufferQueue& GetBufferQueue() {
+ return buffer_queue;
+ }
+
+ /// Gets a const reference to the buffer queue this layer is using.
+ const NVFlinger::BufferQueue& GetBufferQueue() const {
+ return buffer_queue;
+ }
+
+private:
u64 id;
- std::shared_ptr<NVFlinger::BufferQueue> buffer_queue;
+ NVFlinger::BufferQueue& buffer_queue;
};
} // namespace Service::VI
diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp
index a317a2885..74384a24d 100644
--- a/src/core/hle/service/vi/vi.cpp
+++ b/src/core/hle/service/vi/vi.cpp
@@ -525,7 +525,7 @@ private:
LOG_DEBUG(Service_VI, "called. id=0x{:08X} transaction={:X}, flags=0x{:08X}", id,
static_cast<u32>(transaction), flags);
- auto buffer_queue = nv_flinger->FindBufferQueue(id);
+ auto& buffer_queue = nv_flinger->FindBufferQueue(id);
if (transaction == TransactionId::Connect) {
IGBPConnectRequestParcel request{ctx.ReadBuffer()};
@@ -538,7 +538,7 @@ private:
} else if (transaction == TransactionId::SetPreallocatedBuffer) {
IGBPSetPreallocatedBufferRequestParcel request{ctx.ReadBuffer()};
- buffer_queue->SetPreallocatedBuffer(request.data.slot, request.buffer);
+ buffer_queue.SetPreallocatedBuffer(request.data.slot, request.buffer);
IGBPSetPreallocatedBufferResponseParcel response{};
ctx.WriteBuffer(response.Serialize());
@@ -546,7 +546,7 @@ private:
IGBPDequeueBufferRequestParcel request{ctx.ReadBuffer()};
const u32 width{request.data.width};
const u32 height{request.data.height};
- std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
+ std::optional<u32> slot = buffer_queue.DequeueBuffer(width, height);
if (slot) {
// Buffer is available
@@ -559,8 +559,8 @@ private:
[=](Kernel::SharedPtr<Kernel::Thread> thread, Kernel::HLERequestContext& ctx,
Kernel::ThreadWakeupReason reason) {
// Repeat TransactParcel DequeueBuffer when a buffer is available
- auto buffer_queue = nv_flinger->FindBufferQueue(id);
- std::optional<u32> slot = buffer_queue->DequeueBuffer(width, height);
+ auto& buffer_queue = nv_flinger->FindBufferQueue(id);
+ std::optional<u32> slot = buffer_queue.DequeueBuffer(width, height);
ASSERT_MSG(slot != std::nullopt, "Could not dequeue buffer.");
IGBPDequeueBufferResponseParcel response{*slot};
@@ -568,28 +568,28 @@ private:
IPC::ResponseBuilder rb{ctx, 2};
rb.Push(RESULT_SUCCESS);
},
- buffer_queue->GetWritableBufferWaitEvent());
+ buffer_queue.GetWritableBufferWaitEvent());
}
} else if (transaction == TransactionId::RequestBuffer) {
IGBPRequestBufferRequestParcel request{ctx.ReadBuffer()};
- auto& buffer = buffer_queue->RequestBuffer(request.slot);
+ auto& buffer = buffer_queue.RequestBuffer(request.slot);
IGBPRequestBufferResponseParcel response{buffer};
ctx.WriteBuffer(response.Serialize());
} else if (transaction == TransactionId::QueueBuffer) {
IGBPQueueBufferRequestParcel request{ctx.ReadBuffer()};
- buffer_queue->QueueBuffer(request.data.slot, request.data.transform,
- request.data.GetCropRect());
+ buffer_queue.QueueBuffer(request.data.slot, request.data.transform,
+ request.data.GetCropRect());
IGBPQueueBufferResponseParcel response{1280, 720};
ctx.WriteBuffer(response.Serialize());
} else if (transaction == TransactionId::Query) {
IGBPQueryRequestParcel request{ctx.ReadBuffer()};
- u32 value =
- buffer_queue->Query(static_cast<NVFlinger::BufferQueue::QueryType>(request.type));
+ const u32 value =
+ buffer_queue.Query(static_cast<NVFlinger::BufferQueue::QueryType>(request.type));
IGBPQueryResponseParcel response{value};
ctx.WriteBuffer(response.Serialize());
@@ -629,12 +629,12 @@ private:
LOG_WARNING(Service_VI, "(STUBBED) called id={}, unknown={:08X}", id, unknown);
- const auto buffer_queue = nv_flinger->FindBufferQueue(id);
+ const auto& buffer_queue = nv_flinger->FindBufferQueue(id);
// TODO(Subv): Find out what this actually is.
IPC::ResponseBuilder rb{ctx, 2, 1};
rb.Push(RESULT_SUCCESS);
- rb.PushCopyObjects(buffer_queue->GetBufferWaitEvent());
+ rb.PushCopyObjects(buffer_queue.GetBufferWaitEvent());
}
std::shared_ptr<NVFlinger::NVFlinger> nv_flinger;
@@ -752,6 +752,7 @@ public:
{1102, nullptr, "GetDisplayResolution"},
{2010, &IManagerDisplayService::CreateManagedLayer, "CreateManagedLayer"},
{2011, nullptr, "DestroyManagedLayer"},
+ {2012, nullptr, "CreateStrayLayer"},
{2050, nullptr, "CreateIndirectLayer"},
{2051, nullptr, "DestroyIndirectLayer"},
{2052, nullptr, "CreateIndirectProducerEndPoint"},
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index 1f425f90b..252592edd 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -376,9 +376,9 @@ enum class R2pMode : u64 {
};
enum class IpaInterpMode : u64 {
- Linear = 0,
- Perspective = 1,
- Flat = 2,
+ Pass = 0,
+ Multiply = 1,
+ Constant = 2,
Sc = 3,
};
diff --git a/src/video_core/engines/shader_header.h b/src/video_core/engines/shader_header.h
index cf2b76ff6..e86a7f04a 100644
--- a/src/video_core/engines/shader_header.h
+++ b/src/video_core/engines/shader_header.h
@@ -16,6 +16,13 @@ enum class OutputTopology : u32 {
TriangleStrip = 7,
};
+enum class AttributeUse : u8 {
+ Unused = 0,
+ Constant = 1,
+ Perspective = 2,
+ ScreenLinear = 3,
+};
+
// Documentation in:
// http://download.nvidia.com/open-gpu-doc/Shader-Program-Header/1/Shader-Program-Header.html#ImapTexture
struct Header {
@@ -84,9 +91,15 @@ struct Header {
} vtg;
struct {
- INSERT_PADDING_BYTES(3); // ImapSystemValuesA
- INSERT_PADDING_BYTES(1); // ImapSystemValuesB
- INSERT_PADDING_BYTES(32); // ImapGenericVector[32]
+ INSERT_PADDING_BYTES(3); // ImapSystemValuesA
+ INSERT_PADDING_BYTES(1); // ImapSystemValuesB
+ union {
+ BitField<0, 2, AttributeUse> x;
+ BitField<2, 2, AttributeUse> y;
+ BitField<4, 2, AttributeUse> w;
+ BitField<6, 2, AttributeUse> z;
+ u8 raw;
+ } imap_generic_vector[32];
INSERT_PADDING_BYTES(2); // ImapColor
INSERT_PADDING_BYTES(2); // ImapSystemValuesC
INSERT_PADDING_BYTES(10); // ImapFixedFncTexture[10]
@@ -103,6 +116,28 @@ struct Header {
const u32 bit = render_target * 4 + component;
return omap.target & (1 << bit);
}
+ AttributeUse GetAttributeIndexUse(u32 attribute, u32 index) const {
+ return static_cast<AttributeUse>(
+ (imap_generic_vector[attribute].raw >> (index * 2)) & 0x03);
+ }
+ AttributeUse GetAttributeUse(u32 attribute) const {
+ AttributeUse result = AttributeUse::Unused;
+ for (u32 i = 0; i < 4; i++) {
+ const auto index = GetAttributeIndexUse(attribute, i);
+ if (index == AttributeUse::Unused) {
+ continue;
+ }
+ if (result == AttributeUse::Unused || result == index) {
+ result = index;
+ continue;
+ }
+ LOG_CRITICAL(HW_GPU, "Generic Attribute Conflict in Interpolation Mode");
+ if (index == AttributeUse::Perspective) {
+ result = index;
+ }
+ }
+ return result;
+ }
} ps;
};
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index 74200914e..e6d47ce41 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -1257,7 +1257,11 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface,
case SurfaceTarget::TextureCubemap:
case SurfaceTarget::Texture2DArray:
case SurfaceTarget::TextureCubeArray:
- FastLayeredCopySurface(old_surface, new_surface);
+ if (old_params.pixel_format == new_params.pixel_format)
+ FastLayeredCopySurface(old_surface, new_surface);
+ else {
+ AccurateCopySurface(old_surface, new_surface);
+ }
break;
default:
LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}",
diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
index db18f4dbe..72ff6ac6a 100644
--- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp
@@ -20,6 +20,7 @@
namespace OpenGL::GLShader {
using Tegra::Shader::Attribute;
+using Tegra::Shader::AttributeUse;
using Tegra::Shader::Header;
using Tegra::Shader::IpaInterpMode;
using Tegra::Shader::IpaMode;
@@ -288,34 +289,22 @@ private:
code.AddNewLine();
}
- std::string GetInputFlags(const IpaMode& input_mode) {
- const IpaSampleMode sample_mode = input_mode.sampling_mode;
- const IpaInterpMode interp_mode = input_mode.interpolation_mode;
+ std::string GetInputFlags(AttributeUse attribute) {
std::string out;
- switch (interp_mode) {
- case IpaInterpMode::Flat:
+ switch (attribute) {
+ case AttributeUse::Constant:
out += "flat ";
break;
- case IpaInterpMode::Linear:
+ case AttributeUse::ScreenLinear:
out += "noperspective ";
break;
- case IpaInterpMode::Perspective:
+ case AttributeUse::Perspective:
// Default, Smooth
break;
default:
- UNIMPLEMENTED_MSG("Unhandled IPA interp mode: {}", static_cast<u32>(interp_mode));
- }
- switch (sample_mode) {
- case IpaSampleMode::Centroid:
- // It can be implemented with the "centroid " keyword in GLSL
- UNIMPLEMENTED_MSG("Unimplemented IPA sampler mode centroid");
- break;
- case IpaSampleMode::Default:
- // Default, n/a
- break;
- default:
- UNIMPLEMENTED_MSG("Unimplemented IPA sampler mode: {}", static_cast<u32>(sample_mode));
+ LOG_CRITICAL(HW_GPU, "Unused attribute being fetched");
+ UNREACHABLE();
}
return out;
}
@@ -324,16 +313,11 @@ private:
const auto& attributes = ir.GetInputAttributes();
for (const auto element : attributes) {
const Attribute::Index index = element.first;
- const IpaMode& input_mode = *element.second.begin();
if (index < Attribute::Index::Attribute_0 || index > Attribute::Index::Attribute_31) {
// Skip when it's not a generic attribute
continue;
}
- ASSERT(element.second.size() > 0);
- UNIMPLEMENTED_IF_MSG(element.second.size() > 1,
- "Multiple input flag modes are not supported in GLSL");
-
// TODO(bunnei): Use proper number of elements for these
u32 idx = static_cast<u32>(index) - static_cast<u32>(Attribute::Index::Attribute_0);
if (stage != ShaderStage::Vertex) {
@@ -345,8 +329,14 @@ private:
if (stage == ShaderStage::Geometry) {
attr = "gs_" + attr + "[]";
}
- code.AddLine("layout (location = " + std::to_string(idx) + ") " +
- GetInputFlags(input_mode) + "in vec4 " + attr + ';');
+ std::string suffix;
+ if (stage == ShaderStage::Fragment) {
+ const auto input_mode =
+ header.ps.GetAttributeUse(idx - GENERIC_VARYING_START_LOCATION);
+ suffix = GetInputFlags(input_mode);
+ }
+ code.AddLine("layout (location = " + std::to_string(idx) + ") " + suffix + "in vec4 " +
+ attr + ';');
}
if (!attributes.empty())
code.AddNewLine();
@@ -1584,4 +1574,4 @@ ProgramResult Decompile(const ShaderIR& ir, Maxwell::ShaderStage stage, const st
return {decompiler.GetResult(), decompiler.GetShaderEntries()};
}
-} // namespace OpenGL::GLShader \ No newline at end of file
+} // namespace OpenGL::GLShader
diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp
index 04e1db911..7d96649af 100644
--- a/src/video_core/renderer_opengl/gl_shader_gen.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp
@@ -124,7 +124,7 @@ layout (location = 5) out vec4 FragColor5;
layout (location = 6) out vec4 FragColor6;
layout (location = 7) out vec4 FragColor7;
-layout (location = 0) in vec4 position;
+layout (location = 0) in noperspective vec4 position;
layout (std140, binding = EMULATION_UBO_BINDING) uniform fs_config {
vec4 viewport_flip;
@@ -172,4 +172,4 @@ void main() {
return {out, program.second};
}
-} // namespace OpenGL::GLShader \ No newline at end of file
+} // namespace OpenGL::GLShader
diff --git a/src/video_core/shader/decode/memory.cpp b/src/video_core/shader/decode/memory.cpp
index 55ec601ff..38f01ca50 100644
--- a/src/video_core/shader/decode/memory.cpp
+++ b/src/video_core/shader/decode/memory.cpp
@@ -48,7 +48,7 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) {
UNIMPLEMENTED_IF_MSG((instr.attribute.fmt20.immediate.Value() % sizeof(u32)) != 0,
"Unaligned attribute loads are not supported");
- Tegra::Shader::IpaMode input_mode{Tegra::Shader::IpaInterpMode::Perspective,
+ Tegra::Shader::IpaMode input_mode{Tegra::Shader::IpaInterpMode::Pass,
Tegra::Shader::IpaSampleMode::Default};
u64 next_element = instr.attribute.fmt20.element;
diff --git a/src/video_core/shader/decode/other.cpp b/src/video_core/shader/decode/other.cpp
index f9502e3d0..d750a2936 100644
--- a/src/video_core/shader/decode/other.cpp
+++ b/src/video_core/shader/decode/other.cpp
@@ -135,7 +135,18 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
instr.ipa.sample_mode.Value()};
const Node attr = GetInputAttribute(attribute.index, attribute.element, input_mode);
- const Node value = GetSaturatedFloat(attr, instr.ipa.saturate);
+ Node value = attr;
+ const Tegra::Shader::Attribute::Index index = attribute.index.Value();
+ if (index >= Tegra::Shader::Attribute::Index::Attribute_0 &&
+ index <= Tegra::Shader::Attribute::Index::Attribute_31) {
+ // TODO(Blinkhawk): There are cases where a perspective attribute use PASS.
+ // In theory by setting them as perspective, OpenGL does the perspective correction.
+ // A way must figured to reverse the last step of it.
+ if (input_mode.interpolation_mode == Tegra::Shader::IpaInterpMode::Multiply) {
+ value = Operation(OperationCode::FMul, PRECISE, value, GetRegister(instr.gpr20));
+ }
+ }
+ value = GetSaturatedFloat(value, instr.ipa.saturate);
SetRegister(bb, instr.gpr0, value);
break;
@@ -175,4 +186,4 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) {
return pc;
}
-} // namespace VideoCommon::Shader \ No newline at end of file
+} // namespace VideoCommon::Shader
diff --git a/src/video_core/shader/track.cpp b/src/video_core/shader/track.cpp
index be4635342..33b071747 100644
--- a/src/video_core/shader/track.cpp
+++ b/src/video_core/shader/track.cpp
@@ -20,9 +20,9 @@ std::pair<Node, s64> FindOperation(const NodeBlock& code, s64 cursor,
return {node, cursor};
}
if (const auto conditional = std::get_if<ConditionalNode>(node)) {
- const auto& code = conditional->GetCode();
- const auto [found, internal_cursor] =
- FindOperation(code, static_cast<s64>(code.size() - 1), operation_code);
+ const auto& conditional_code = conditional->GetCode();
+ const auto [found, internal_cursor] = FindOperation(
+ conditional_code, static_cast<s64>(conditional_code.size() - 1), operation_code);
if (found)
return {found, cursor};
}
@@ -58,8 +58,8 @@ Node ShaderIR::TrackCbuf(Node tracked, const NodeBlock& code, s64 cursor) {
return nullptr;
}
if (const auto conditional = std::get_if<ConditionalNode>(tracked)) {
- const auto& code = conditional->GetCode();
- return TrackCbuf(tracked, code, static_cast<s64>(code.size()));
+ const auto& conditional_code = conditional->GetCode();
+ return TrackCbuf(tracked, conditional_code, static_cast<s64>(conditional_code.size()));
}
return nullptr;
}
generated by cgit v1.2.3 (git 2.25.1) at 2025-10-31 15:26:31 +0000