Merge branch 'master' of https://github.com/Arisotura/melonDS into sw…

…itch-new

.github/workflows/build-ubuntu-aarch64.yml

@@ -20,11 +20,9 @@ jobs:
     - name: Upgrade system
       shell: bash
       working-directory: ${{runner.workspace}}
-      run: | #Fix grub installation error - https://github.com/actions/virtual-environments/issues/1605
+      run: |
         sudo apt update
-        sudo apt-get install grub-efi  
-        sudo update-grub
-        sudo apt full-upgrade
+        sudo ACCEPT_EULA=Y apt full-upgrade
     - name: Install dependencies
       shell: bash
       working-directory: ${{runner.workspace}}

README.md

@@ -5,6 +5,11 @@
 <a href="http://melonds.kuribo64.net/downloads.php" alt="Release: 0.9.1"><img src="https://img.shields.io/badge/release-0.9.1-%235c913b.svg"></a>
 <a href="https://www.gnu.org/licenses/gpl-3.0" alt="License: GPLv3"><img src="https://img.shields.io/badge/License-GPL%20v3-%23ff554d.svg"></a>
 <a href="https://kiwiirc.com/client/irc.badnik.net/?nick=IRC-Source_?#melonds" alt="IRC channel: #melonds"><img src="https://img.shields.io/badge/IRC%20chat-%23melonds-%23dd2e44.svg"></a>
+<br>
+<a href="https://github.com/Arisotura/melonDS/actions?query=workflow%3A%22CMake+Build+%28Windows+x86-64%29%22+event%3Apush"><img src="https://img.shields.io/github/workflow/status/Arisotura/melonDS/CMake%20Build%20(Windows%20x86-64)?label=Windows%20x86-64&logo=GitHub"></img></a>
+<a href="https://github.com/Arisotura/melonDS/actions?query=workflow%3A%22CMake+Build+%28Ubuntu+x86-64%29%22+event%3Apush"><img src="https://img.shields.io/github/workflow/status/Arisotura/melonDS/CMake%20Build%20(Ubuntu%20x86-64)?label=Linux%20x86-64&logo=GitHub"></img></a>
+<a href="https://github.com/Arisotura/melonDS/actions?query=workflow%3A%22CMake+Build+%28macOS+x86-64%29%22+event%3Apush"><img src="https://img.shields.io/github/workflow/status/Arisotura/melonDS/CMake%20Build%20(macOS%20x86-64)?label=macOS%20x86-64&logo=GitHub"></img></a>
+<a href="https://github.com/Arisotura/melonDS/actions?query=workflow%3A%22CMake+Build+%28Ubuntu+aarch64%29%22+event%3Apush"><img src="https://img.shields.io/github/workflow/status/Arisotura/melonDS/CMake%20Build%20(Ubuntu%20aarch64)?label=Linux%20ARM64&logo=GitHub"></img></a>
 </p>
 DS emulator, sorta
 

src/ARMJIT_Memory.cpp

@@ -169,6 +169,8 @@ static void SigsegvHandler(int sig, siginfo_t* info, void* rawContext)
         desc.FaultPC = (u8*)context->uc_mcontext->__ss.__rip;
     #elif defined(__FreeBSD__)
         desc.FaultPC = (u8*)context->uc_mcontext.mc_rip;
+    #elif defined(__NetBSD__)
+        desc.FaultPC = (u8*)context->uc_mcontext.__gregs[_REG_RIP];
     #else
         desc.FaultPC = (u8*)context->uc_mcontext.gregs[REG_RIP];
     #endif
@@ -190,6 +192,8 @@ static void SigsegvHandler(int sig, siginfo_t* info, void* rawContext)
             context->uc_mcontext->__ss.__rip = (u64)desc.FaultPC;
         #elif defined(__FreeBSD__)
             context->uc_mcontext.mc_rip = (u64)desc.FaultPC;
+        #elif defined(__NetBSD__)
+            context->uc_mcontext.__gregs[_REG_RIP] = (u64)desc.FaultPC;
         #else
             context->uc_mcontext.gregs[REG_RIP] = (u64)desc.FaultPC;
         #endif
@@ -717,16 +721,17 @@ void Init()
 
     MemoryFile = CreateFileMapping(INVALID_HANDLE_VALUE, NULL, PAGE_READWRITE, 0, MemoryTotalSize, NULL);
 
-    MemoryBase = (u8*)VirtualAlloc(NULL, MemoryTotalSize, MEM_RESERVE, PAGE_READWRITE);
-
-    FastMem9Start = VirtualAlloc(NULL, AddrSpaceSize, MEM_RESERVE, PAGE_READWRITE);
-    FastMem7Start = VirtualAlloc(NULL, AddrSpaceSize, MEM_RESERVE, PAGE_READWRITE);
-
-    // only free them after they have all been reserved
-    // so they can't overlap
+    MemoryBase = (u8*)VirtualAlloc(NULL, AddrSpaceSize*4, MEM_RESERVE, PAGE_READWRITE);
     VirtualFree(MemoryBase, 0, MEM_RELEASE);
-    VirtualFree(FastMem9Start, 0, MEM_RELEASE);
-    VirtualFree(FastMem7Start, 0, MEM_RELEASE);
+    // this is incredible hacky
+    // but someone else is trying to go into our address space!
+    // Windows will very likely give them virtual memory starting at the same address
+    // as it is giving us now.
+    // That's why we don't use this address, but instead 4gb inwards
+    // I know this is terrible
+    FastMem9Start = MemoryBase + AddrSpaceSize;
+    FastMem7Start = MemoryBase + AddrSpaceSize*2;
+    MemoryBase = MemoryBase + AddrSpaceSize*3;
 
     MapViewOfFileEx(MemoryFile, FILE_MAP_READ | FILE_MAP_WRITE, 0, 0, MemoryTotalSize, MemoryBase);
 

src/CMakeLists.txt

@@ -1,6 +1,6 @@
 project(core)
 
-set (CMAKE_CXX_STANDARD 14)
+set (CMAKE_CXX_STANDARD 17)
 
 add_library(core STATIC
 	ARCodeFile.cpp

src/GPU.cpp

@@ -21,6 +21,7 @@
 #include "NDS.h"
 #include "GPU.h"
 
+#include "GPU2D_Soft.h"
 
 namespace GPU
 {
@@ -78,11 +79,10 @@ u8* VRAMPtr_BOBJ[0x8];
 
 int FrontBuffer;
 u32* Framebuffer[2][2];
-int Renderer;
-bool Accelerated;
+int Renderer = 0;
 
-GPU2D* GPU2D_A;
-GPU2D* GPU2D_B;
+std::unique_ptr<GPU2D> GPU2D_A = {};
+std::unique_ptr<GPU2D> GPU2D_B = {};
 
 /*
     VRAM invalidation tracking
@@ -141,30 +141,33 @@ u32 PaletteDirty;
 
 u8 AllPaletteMemory[2*1024+BGExtPalSize*2+OBJExtPalSize*2];
 
+#ifdef OGLRENDERER_ENABLED
+std::unique_ptr<GLCompositor> CurGLCompositor = {};
+#endif
+
 bool Init()
 {
-#ifndef NEONSOFTGPU_ENABLED
-    GPU2D_A = new GPU2D_Soft(0);
-    GPU2D_B = new GPU2D_Soft(1);
+#ifdef NEONGPU_ENABLED
+    GPU2D_A = std::make_unique<GPU2D_Soft>(0);
+    GPU2D_B = std::make_unique<GPU2D_Soft>(1);
 #else
-    GPU2D_A = new GPU2D_NeonSoft(0);
-    GPU2D_B = new GPU2D_NeonSoft(1);
+    GPU2D_A = std::make_unique<GPU2D_NeonSoft>(0);
+    GPU2D_B = std::make_unique<GPU2D_NeonSoft>(1);
 #endif
     if (!GPU3D::Init()) return false;
 
     FrontBuffer = 0;
     Framebuffer[0][0] = NULL; Framebuffer[0][1] = NULL;
     Framebuffer[1][0] = NULL; Framebuffer[1][1] = NULL;
     Renderer = 0;
-    Accelerated = false;
 
     return true;
 }
 
 void DeInit()
 {
-    delete GPU2D_A;
-    delete GPU2D_B;
+    GPU2D_A.reset();
+    GPU2D_B.reset();
     GPU3D::DeInit();
 
     if (Framebuffer[0][0]) delete[] Framebuffer[0][0];
@@ -254,9 +257,12 @@ void Reset()
     memset(VRAMPtr_BBG, 0, sizeof(VRAMPtr_BBG));
     memset(VRAMPtr_BOBJ, 0, sizeof(VRAMPtr_BOBJ));
 
-    int fbsize;
-    if (Accelerated) fbsize = (256*3 + 1) * 192;
-    else             fbsize = 256 * 192;
+    size_t fbsize;
+    if (GPU3D::CurrentRenderer->Accelerated)
+        fbsize = (256*3 + 1) * 192;
+    else
+        fbsize = 256 * 192;
+
     for (int i = 0; i < fbsize; i++)
     {
         Framebuffer[0][0][i] = 0xFFFFFFFF;
@@ -287,17 +293,22 @@ void Reset()
 void Stop()
 {
     int fbsize;
-    if (Accelerated) fbsize = (256*3 + 1) * 192;
-    else             fbsize = 256 * 192;
+    if (GPU3D::CurrentRenderer->Accelerated)
+        fbsize = (256*3 + 1) * 192;
+    else
+        fbsize = 256 * 192;
+
     memset(Framebuffer[0][0], 0, fbsize*4);
     memset(Framebuffer[0][1], 0, fbsize*4);
     memset(Framebuffer[1][0], 0, fbsize*4);
     memset(Framebuffer[1][1], 0, fbsize*4);
 
 #ifdef OGLRENDERER_ENABLED
-    if (Accelerated)
-        GLCompositor::Stop(); 
-#endif  
+    // This needs a better way to know that we're
+    // using the OpenGL renderer specifically
+    if (GPU3D::CurrentRenderer->Accelerated)
+        CurGLCompositor->Stop();
+#endif
 }
 
 void DoSavestate(Savestate* file)
@@ -386,37 +397,42 @@ void InitRenderer(int renderer)
 #ifdef OGLRENDERER_ENABLED
     if (renderer == 1)
     {
-        if (!GLCompositor::Init())
+        CurGLCompositor = std::make_unique<GLCompositor>();
+        // Create opengl rendrerer
+        if (!CurGLCompositor->Init())
         {
+            // Fallback on software renderer
             renderer = 0;
+            GPU3D::CurrentRenderer = std::make_unique<GPU3D::SoftRenderer>();
+            GPU3D::CurrentRenderer->Init();
         }
-        else if (!GPU3D::GLRenderer::Init())
+        GPU3D::CurrentRenderer = std::make_unique<GPU3D::GLRenderer>();
+        if (!GPU3D::CurrentRenderer->Init())
         {
-            GLCompositor::DeInit();
+            // Fallback on software renderer
+            CurGLCompositor->DeInit();
+            CurGLCompositor.reset();
             renderer = 0;
+            GPU3D::CurrentRenderer = std::make_unique<GPU3D::SoftRenderer>();
         }
     }
     else
 #endif
     {
-        GPU3D::SoftRenderer::Init();
+        GPU3D::CurrentRenderer = std::make_unique<GPU3D::SoftRenderer>();
+        GPU3D::CurrentRenderer->Init();
     }
 
     Renderer = renderer;
-    Accelerated = renderer != 0;
 }
 
 void DeInitRenderer()
 {
-    if (Renderer == 0)
-    {
-        GPU3D::SoftRenderer::DeInit();
-    }
+    GPU3D::CurrentRenderer->DeInit();
 #ifdef OGLRENDERER_ENABLED
-    else
+    if (Renderer == 1)
     {
-        GPU3D::GLRenderer::DeInit();
-        GLCompositor::DeInit();
+        CurGLCompositor->DeInit();
     }
 #endif
 }
@@ -425,13 +441,13 @@ void ResetRenderer()
 {
     if (Renderer == 0)
     {
-        GPU3D::SoftRenderer::Reset();
+        GPU3D::CurrentRenderer->Reset();
     }
 #ifdef OGLRENDERER_ENABLED
     else
     {
-        GLCompositor::Reset();
-        GPU3D::GLRenderer::Reset();
+        CurGLCompositor->Reset();
+        GPU3D::CurrentRenderer->Reset();
     }
 #endif
 }
@@ -444,10 +460,12 @@ void SetRenderSettings(int renderer, RenderSettings& settings)
         InitRenderer(renderer);
     }
 
-    bool accel = Accelerated;
     int fbsize;
-    if (accel) fbsize = (256*3 + 1) * 192;
-    else       fbsize = 256 * 192;
+    if (GPU3D::CurrentRenderer->Accelerated)
+        fbsize = (256*3 + 1) * 192;
+    else
+        fbsize = 256 * 192;
+
     if (Framebuffer[0][0]) { delete[] Framebuffer[0][0]; Framebuffer[0][0] = nullptr; }
     if (Framebuffer[1][0]) { delete[] Framebuffer[1][0]; Framebuffer[1][0] = nullptr; }
     if (Framebuffer[0][1]) { delete[] Framebuffer[0][1]; Framebuffer[0][1] = nullptr; }
@@ -465,18 +483,15 @@ void SetRenderSettings(int renderer, RenderSettings& settings)
 
     AssignFramebuffers();
 
-    GPU2D_A->SetRenderSettings(accel);
-    GPU2D_B->SetRenderSettings(accel);
-
     if (Renderer == 0)
     {
-        GPU3D::SoftRenderer::SetRenderSettings(settings);
+        GPU3D::CurrentRenderer->SetRenderSettings(settings);
     }
 #ifdef OGLRENDERER_ENABLED
     else
     {
-        GLCompositor::SetRenderSettings(settings);
-        GPU3D::GLRenderer::SetRenderSettings(settings);
+        CurGLCompositor->SetRenderSettings(settings);
+        GPU3D::CurrentRenderer->SetRenderSettings(settings);
     }
 #endif
 }
@@ -1153,7 +1168,9 @@ void StartScanline(u32 line)
             GPU3D::VBlank();
 
 #ifdef OGLRENDERER_ENABLED
-            if (Accelerated) GLCompositor::RenderFrame();
+            // Need a better way to identify the openGL renderer in particular
+            if (GPU3D::CurrentRenderer->Accelerated)
+                CurGLCompositor->RenderFrame();
 #endif
         }
     }
@@ -1191,7 +1208,7 @@ NonStupidBitField<Size/VRAMDirtyGranularity> VRAMTrackingSet<Size, MappingGranul
     {
         if (currentMappings[i] != Mapping[i])
         {
-            result |= NonStupidBitField<Size/VRAMDirtyGranularity>(i*VRAMBitsPerMapping, VRAMBitsPerMapping);
+            result.SetRange(i*VRAMBitsPerMapping, VRAMBitsPerMapping);
             banksToBeZeroed |= currentMappings[i];
             Mapping[i] = currentMappings[i];
         }
@@ -1213,19 +1230,19 @@ NonStupidBitField<Size/VRAMDirtyGranularity> VRAMTrackingSet<Size, MappingGranul
                 if (MappingGranularity == 16*1024)
                 {
                     u32 dirty = ((u32*)VRAMDirty[num].Data)[i & (VRAMMask[num] >> 14)];
-                    ((u32*)result.Data)[i] |= dirty;
+                    result.Data[i / 2] |= (u64)dirty << ((i&1)*32);
                 }
                 else if (MappingGranularity == 8*1024)
                 {
                     u16 dirty = ((u16*)VRAMDirty[num].Data)[i & (VRAMMask[num] >> 13)];
-                    ((u16*)result.Data)[i] |= dirty;
+                    result.Data[i / 4] |= (u64)dirty << ((i&3)*16);
                 }
                 else if (MappingGranularity == 128*1024)
                 {
-                    ((u64*)result.Data)[i * 4 + 0] |= ((u64*)VRAMDirty[num].Data)[0];
-                    ((u64*)result.Data)[i * 4 + 1] |= ((u64*)VRAMDirty[num].Data)[1];
-                    ((u64*)result.Data)[i * 4 + 2] |= ((u64*)VRAMDirty[num].Data)[2];
-                    ((u64*)result.Data)[i * 4 + 3] |= ((u64*)VRAMDirty[num].Data)[3];
+                    result.Data[i * 4 + 0] |= VRAMDirty[num].Data[0];
+                    result.Data[i * 4 + 1] |= VRAMDirty[num].Data[1];
+                    result.Data[i * 4 + 2] |= VRAMDirty[num].Data[2];
+                    result.Data[i * 4 + 3] |= VRAMDirty[num].Data[3];
                 }
                 else
                 {
@@ -1240,7 +1257,7 @@ NonStupidBitField<Size/VRAMDirtyGranularity> VRAMTrackingSet<Size, MappingGranul
     {
         u32 num = __builtin_ctz(banksToBeZeroed);
         banksToBeZeroed &= ~(1 << num);
-        memset(VRAMDirty[num].Data, 0, sizeof(VRAMDirty[num].Data));
+        VRAMDirty[num].Clear();
     }
 
     return result;
@@ -1270,7 +1287,7 @@ void SyncDirtyFlags(u32* mappings, NonStupidBitField<Size>& writtenFlags)
             mapping &= ~(1 << num);
         }
     }
-    memset(writtenFlags.Data, 0, sizeof(writtenFlags.Data));
+    writtenFlags.Clear();
 }
 
 void SyncDirtyFlags()