From 345e73f2feb0701e3c3099d002a1c21fb524eae4 Mon Sep 17 00:00:00 2001
From: ReinUsesLisp <reinuseslisp@airmail.cc>
Date: Fri, 10 May 2019 04:17:48 -0300
Subject: video_core: Use un-shifted block sizes to avoid integer divisions

Instead of storing all block width, height and depths in their shifted
form:

block_width = 1U << block_shift;

Store them like they are provided by the emulated hardware (their
block_shift form). This way we can avoid doing the costly
Common::AlignUp operation to align texture sizes and drop CPU integer
divisions with bitwise logic (defined in Common::AlignBits).
---
 src/video_core/engines/fermi_2d.h | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

(limited to 'src/video_core/engines/fermi_2d.h')

diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h
index 45f59a4d9..3d28afa91 100644
--- a/src/video_core/engines/fermi_2d.h
+++ b/src/video_core/engines/fermi_2d.h
@@ -63,18 +63,15 @@ public:
             }
 
             u32 BlockWidth() const {
-                // The block width is stored in log2 format.
-                return 1 << block_width;
+                return block_width;
             }
 
             u32 BlockHeight() const {
-                // The block height is stored in log2 format.
-                return 1 << block_height;
+                return block_height;
             }
 
             u32 BlockDepth() const {
-                // The block depth is stored in log2 format.
-                return 1 << block_depth;
+                return block_depth;
             }
         };
         static_assert(sizeof(Surface) == 0x28, "Surface has incorrect size");
-- 
cgit v1.2.3


From 175aa343ff1c9f931b266caf2d19b8df943dab0d Mon Sep 17 00:00:00 2001
From: Fernando Sahmkow <fsahmkow27@gmail.com>
Date: Sat, 18 May 2019 04:57:49 -0400
Subject: texture_cache: Fermi2D reform and implement View Mirage

This also does some fixes on compressed textures reinterpret and on the
Fermi2D engine in general.
---
 src/video_core/engines/fermi_2d.h | 44 +++++++++++++++++++++++++++++----------
 1 file changed, 33 insertions(+), 11 deletions(-)

(limited to 'src/video_core/engines/fermi_2d.h')

diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h
index 3d28afa91..0a4c7c5ad 100644
--- a/src/video_core/engines/fermi_2d.h
+++ b/src/video_core/engines/fermi_2d.h
@@ -9,6 +9,7 @@
 #include "common/bit_field.h"
 #include "common/common_funcs.h"
 #include "common/common_types.h"
+#include "common/math_util.h"
 #include "video_core/gpu.h"
 
 namespace Tegra {
@@ -38,6 +39,26 @@ public:
     /// Write the value to the register identified by method.
     void CallMethod(const GPU::MethodCall& method_call);
 
+    enum class Origin : u32 {
+        Center = 0,
+        Corner = 1,
+    };
+
+    enum class Filter : u32 {
+        PointSample = 0, // Nearest
+        Linear = 1,
+    };
+
+    enum class Operation : u32 {
+        SrcCopyAnd = 0,
+        ROPAnd = 1,
+        Blend = 2,
+        SrcCopy = 3,
+        ROP = 4,
+        SrcCopyPremult = 5,
+        BlendPremult = 6,
+    };
+
     struct Regs {
         static constexpr std::size_t NUM_REGS = 0x258;
 
@@ -76,16 +97,6 @@ public:
         };
         static_assert(sizeof(Surface) == 0x28, "Surface has incorrect size");
 
-        enum class Operation : u32 {
-            SrcCopyAnd = 0,
-            ROPAnd = 1,
-            Blend = 2,
-            SrcCopy = 3,
-            ROP = 4,
-            SrcCopyPremult = 5,
-            BlendPremult = 6,
-        };
-
         union {
             struct {
                 INSERT_PADDING_WORDS(0x80);
@@ -102,7 +113,11 @@ public:
 
                 INSERT_PADDING_WORDS(0x177);
 
-                u32 blit_control;
+                union {
+                    u32 raw;
+                    BitField<0, 1, Origin> origin;
+                    BitField<4, 1, Filter> filter;
+                } blit_control;
 
                 INSERT_PADDING_WORDS(0x8);
 
@@ -121,6 +136,13 @@ public:
         };
     } regs{};
 
+    struct Config {
+        Operation operation;
+        Filter filter;
+        Common::Rectangle<u32> src_rect;
+        Common::Rectangle<u32> dst_rect;
+    };
+
 private:
     VideoCore::RasterizerInterface& rasterizer;
     MemoryManager& memory_manager;
-- 
cgit v1.2.3


From 7232a1ed16e46715c29d781fb143bdf799090bec Mon Sep 17 00:00:00 2001
From: Fernando Sahmkow <fsahmkow27@gmail.com>
Date: Thu, 13 Jun 2019 16:41:16 -0400
Subject: decoders: correct block calculation

---
 src/video_core/engines/fermi_2d.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

(limited to 'src/video_core/engines/fermi_2d.h')

diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h
index 0a4c7c5ad..05421d185 100644
--- a/src/video_core/engines/fermi_2d.h
+++ b/src/video_core/engines/fermi_2d.h
@@ -84,15 +84,15 @@ public:
             }
 
             u32 BlockWidth() const {
-                return block_width;
+                return block_width.Value();
             }
 
             u32 BlockHeight() const {
-                return block_height;
+                return block_height.Value();
             }
 
             u32 BlockDepth() const {
-                return block_depth;
+                return block_depth.Value();
             }
         };
         static_assert(sizeof(Surface) == 0x28, "Surface has incorrect size");
-- 
cgit v1.2.3