fabi/Xenia-Canary
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[UI] Image post-processing and full presentation/window rework

Browse Source 
[GPU] Add FXAA post-processing
[UI] Add FidelityFX FSR and CAS post-processing
[UI] Add blue noise dithering from 10bpc to 8bpc
[GPU] Apply the DC PWL gamma ramp closer to the spec, supporting fully white color
[UI] Allow the GPU CP thread to present on the host directly, bypassing the UI thread OS paint event
[UI] Allow variable refresh rate (or tearing)
[UI] Present the newest frame (restart) on DXGI
[UI] Replace GraphicsContext with a far more advanced Presenter with more coherent surface connection and UI overlay state management
[UI] Connect presentation to windows via the Surface class, not native window handles
[Vulkan] Switch to simpler Vulkan setup with no instance/device separation due to interdependencies and to pass fewer objects around
[Vulkan] Lower the minimum required Vulkan version to 1.0
[UI/GPU] Various cleanup, mainly ComPtr usage
[UI] Support per-monitor DPI awareness v2 on Windows
[UI] DPI-scale Dear ImGui
[UI] Replace the remaining non-detachable window delegates with unified window event and input listeners
[UI] Allow listeners to safely destroy or close the window, and to register/unregister listeners without use-after-free and the ABA problem
[UI] Explicit Z ordering of input listeners and UI overlays, top-down for input, bottom-up for drawing
[UI] Add explicit window lifecycle phases
[UI] Replace Window virtual functions with explicit desired state, its application, actual state, its feedback
[UI] GTK: Apply the initial size to the drawing area
[UI] Limit internal UI frame rate to that of the monitor
[UI] Hide the cursor using a timer instead of polling due to no repeated UI thread paints with GPU CP thread presentation, and only within the window
This commit is contained in:
Triang3l
2022-01-29 13:22:03 +03:00
parent 372bdd3ec9
commit fe3f0f26e4
428 changed files with 75942 additions and 18360 deletions
Download Patch File Download Diff File Expand all files Collapse all files

82
src/xenia/gpu/command_processor.h
View File

@@ -2,7 +2,7 @@
******************************************************************************
* Xenia : Xbox 360 Emulator Research Project *
******************************************************************************
* Copyright 2020 Ben Vanik. All rights reserved. *
* Copyright 2022 Ben Vanik. All rights reserved. *
* Released under the BSD license - see LICENSE in the root for more details. *
******************************************************************************
*/
@@ -26,7 +26,7 @@
#include "xenia/gpu/xenos.h"
#include "xenia/kernel/xthread.h"
#include "xenia/memory.h"
#include "xenia/ui/graphics_context.h"
#include "xenia/ui/presenter.h"
namespace xe {
@@ -60,12 +60,20 @@ enum class SwapMode {
enum class GammaRampType {
kUnknown = 0,
kNormal,
kTable,
kPWL,
};
struct GammaRamp {
struct NormalEntry {
// A lot of gamma ramp (DC_LUT) documentation:
// https://developer.amd.com/wordpress/media/2012/10/RRG-216M56-03oOEM.pdf
// The ramps entries are BGR, not RGB.
// For the 256-entry table (used by Direct3D 9 for a 8bpc front buffer),
// 535107D4 has in-game settings allowing separate configuration.
// The component order of the PWL table is untested, however, it's likely BGR
// too, since DC_LUTA/B registers have values for blue first, and for red
// last.
struct TableEntry {
union {
uint32_t value;
struct {
@@ -81,6 +89,15 @@ struct GammaRamp {
union {
uint32_t value;
struct {
// The lower 6 bits are always zero (these are 10-bit in the upper bits
// thus, not fully 16-bit).
// See DC_LUTA/B_CONTROL for information about the way they should be
// interpreted (`output = base + (multiplier * delta) / 2^increment`,
// where the increment is the value specified in DC_LUTA/B_CONTROL for
// the specific color channel, the base is 7 bits of the front buffer
// value above `increment` bits, the multiplier is the lower `increment`
// bits of it; the increment is nonzero, otherwise the 256-entry table
// should be used instead).
uint16_t base;
uint16_t delta;
};
@@ -91,19 +108,25 @@ struct GammaRamp {
union {
PWLValue values[3];
struct {
PWLValue r;
PWLValue g;
PWLValue b;
PWLValue g;
PWLValue r;
};
};
};
NormalEntry normal[256];
TableEntry table[256];
PWLEntry pwl[128];
};
class CommandProcessor {
public:
enum class SwapPostEffect {
kNone,
kFxaa,
kFxaaExtreme,
};
CommandProcessor(GraphicsSystem* graphics_system,
kernel::KernelState* kernel_state);
virtual ~CommandProcessor();
@@ -114,21 +137,26 @@ class CommandProcessor {
Shader* active_vertex_shader() const { return active_vertex_shader_; }
Shader* active_pixel_shader() const { return active_pixel_shader_; }
virtual bool Initialize(std::unique_ptr<xe::ui::GraphicsContext> context);
virtual bool Initialize();
virtual void Shutdown();
void CallInThread(std::function<void()> fn);
virtual void ClearCaches();
SwapState& swap_state() { return swap_state_; }
void set_swap_mode(SwapMode swap_mode) { swap_mode_ = swap_mode; }
void IssueSwap(uint32_t frontbuffer_ptr, uint32_t frontbuffer_width,
uint32_t frontbuffer_height);
void set_swap_request_handler(std::function<void()> fn) {
swap_request_handler_ = fn;
void SetIgnoreSwap(bool ignore_swap) { ignore_swap_ = ignore_swap; }
// "Desired" is for the external thread managing the post-processing effect.
SwapPostEffect GetDesiredSwapPostEffect() const {
return swap_post_effect_desired_;
}
void SetDesiredSwapPostEffect(SwapPostEffect swap_post_effect);
// Implementations must not make assumptions that the front buffer will
// necessarily be a resolve destination - it may be a texture generated by any
// means like written to by the CPU or loaded from a file (the disclaimer
// screen right in the beginning of 4D530AA4 is not a resolved render target,
// for instance).
virtual void IssueSwap(uint32_t frontbuffer_ptr, uint32_t frontbuffer_width,
uint32_t frontbuffer_height) = 0;
// May be called not only from the command processor thread when the command
// processor is paused, and the termination of this function may be explicitly
@@ -179,9 +207,6 @@ class CommandProcessor {
virtual void PrepareForWait();
virtual void ReturnFromWait();
virtual void PerformSwap(uint32_t frontbuffer_ptr, uint32_t frontbuffer_width,
uint32_t frontbuffer_height) = 0;
uint32_t ExecutePrimaryBuffer(uint32_t start_index, uint32_t end_index);
virtual void OnPrimaryBufferEnd() {}
void ExecuteIndirectBuffer(uint32_t ptr, uint32_t length);
@@ -254,6 +279,14 @@ class CommandProcessor {
bool major_mode_explicit) = 0;
virtual bool IssueCopy() = 0;
// "Actual" is for the command processor thread, to be read by the
// implementations.
SwapPostEffect GetActualSwapPostEffect() const {
return swap_post_effect_actual_;
}
// TODO(Triang3l): Write the gamma ramp (including the display controller
// write pointers) in the common code.
virtual void InitializeTrace() = 0;
Memory* memory_ = nullptr;
@@ -274,10 +307,8 @@ class CommandProcessor {
std::atomic<bool> worker_running_;
kernel::object_ref<kernel::XHostThread> worker_thread_;
std::unique_ptr<xe::ui::GraphicsContext> context_;
SwapMode swap_mode_ = SwapMode::kNormal;
SwapState swap_state_;
std::function<void()> swap_request_handler_;
bool ignore_swap_ = false;
std::queue<std::function<void()>> pending_fns_;
// MicroEngine binary from PM4_ME_INIT
@@ -305,8 +336,13 @@ class CommandProcessor {
GammaRamp gamma_ramp_ = {};
int gamma_ramp_rw_subindex_ = 0;
bool dirty_gamma_ramp_normal_ = true;
bool dirty_gamma_ramp_table_ = true;
bool dirty_gamma_ramp_pwl_ = true;
// By default (such as for tools), post-processing is disabled.
// "Desired" is for the external thread managing the post-processing effect.
SwapPostEffect swap_post_effect_desired_ = SwapPostEffect::kNone;
SwapPostEffect swap_post_effect_actual_ = SwapPostEffect::kNone;
};
} // namespace gpu