From 5b2b6d594c6cfa77c3fb92faee63ad524bfe7204 Mon Sep 17 00:00:00 2001
From: ReinUsesLisp <reinuseslisp@airmail.cc>
Date: Thu, 4 Jun 2020 23:03:49 -0300
Subject: shader/texture: Join separate image and sampler pairs offline

Games using D3D idioms can join images and samplers when a shader
executes, instead of baking them into a combined sampler image. This is
also possible on Vulkan.

One approach to this solution would be to use separate samplers on
Vulkan and leave this unimplemented on OpenGL, but we can't do this
because there's no consistent way of determining which constant buffer
holds a sampler and which one an image. We could in theory find the
first bit and if it's in the TIC area, it's an image; but this falls
apart when an image or sampler handle use an index of zero.

The used approach is to track for a LOP.OR operation (this is done at an
IR level, not at an ISA level), track again the constant buffers used as
source and store this pair. Then, outside of shader execution, join
the sample and image pair with a bitwise or operation.

This approach won't work on games that truly use separate samplers in a
meaningful way. For example, pooling textures in a 2D array and
determining at runtime what sampler to use.

This invalidates OpenGL's disk shader cache :)

- Used mostly by D3D ports to Switch
---
 src/video_core/shader/registry.h | 35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)

(limited to 'src/video_core/shader/registry.h')

diff --git a/src/video_core/shader/registry.h b/src/video_core/shader/registry.h
index 0c80d35fd..231206765 100644
--- a/src/video_core/shader/registry.h
+++ b/src/video_core/shader/registry.h
@@ -19,8 +19,39 @@
 
 namespace VideoCommon::Shader {
 
+struct SeparateSamplerKey {
+    std::pair<u32, u32> buffers;
+    std::pair<u32, u32> offsets;
+};
+
+} // namespace VideoCommon::Shader
+
+namespace std {
+
+template <>
+struct hash<VideoCommon::Shader::SeparateSamplerKey> {
+    std::size_t operator()(const VideoCommon::Shader::SeparateSamplerKey& key) const noexcept {
+        return std::hash<u32>{}(key.buffers.first ^ key.buffers.second ^ key.offsets.first ^
+                                key.offsets.second);
+    }
+};
+
+template <>
+struct equal_to<VideoCommon::Shader::SeparateSamplerKey> {
+    bool operator()(const VideoCommon::Shader::SeparateSamplerKey& lhs,
+                    const VideoCommon::Shader::SeparateSamplerKey& rhs) const noexcept {
+        return lhs.buffers == rhs.buffers && lhs.offsets == rhs.offsets;
+    }
+};
+
+} // namespace std
+
+namespace VideoCommon::Shader {
+
 using KeyMap = std::unordered_map<std::pair<u32, u32>, u32, Common::PairHash>;
 using BoundSamplerMap = std::unordered_map<u32, Tegra::Engines::SamplerDescriptor>;
+using SeparateSamplerMap =
+    std::unordered_map<SeparateSamplerKey, Tegra::Engines::SamplerDescriptor>;
 using BindlessSamplerMap =
     std::unordered_map<std::pair<u32, u32>, Tegra::Engines::SamplerDescriptor, Common::PairHash>;
 
@@ -73,6 +104,9 @@ public:
 
     std::optional<Tegra::Engines::SamplerDescriptor> ObtainBoundSampler(u32 offset);
 
+    std::optional<Tegra::Engines::SamplerDescriptor> ObtainSeparateSampler(
+        std::pair<u32, u32> buffers, std::pair<u32, u32> offsets);
+
     std::optional<Tegra::Engines::SamplerDescriptor> ObtainBindlessSampler(u32 buffer, u32 offset);
 
     /// Inserts a key.
@@ -128,6 +162,7 @@ private:
     Tegra::Engines::ConstBufferEngineInterface* engine = nullptr;
     KeyMap keys;
     BoundSamplerMap bound_samplers;
+    SeparateSamplerMap separate_samplers;
     BindlessSamplerMap bindless_samplers;
     u32 bound_buffer;
     GraphicsInfo graphics_info;
-- 
cgit v1.2.3