[video_core] buffer cache readback for nce invalidations

.patch/openssl-cmake/0005-cross-compile.patch

@@ -1,40 +0,0 @@
-diff --git a/cmake/ConfigureOpenSSL.cmake b/cmake/ConfigureOpenSSL.cmake
-index 3012e05..2ae23ff 100644
---- a/cmake/ConfigureOpenSSL.cmake
-+++ b/cmake/ConfigureOpenSSL.cmake
-@@ -108,7 +108,8 @@ function(configure_openssl)
-     )
- 
-     if(NOT "${CONFIGURE_OPTIONS_OLD}" STREQUAL "")
--        if(CONFIGURE_OPTIONS STREQUAL CONFIGURE_OPTIONS_OLD)
-+        # TODO(lizzie): Emscripten has issues with rebuilding due to the wrapper it uses
-+        if(CMAKE_SYSTEM_NAME MATCHES "Emscripten" OR CONFIGURE_OPTIONS STREQUAL CONFIGURE_OPTIONS_OLD)
-             message(STATUS "Found previous configure results. Don't perform configuration")
-             return()
-         endif()
-@@ -134,10 +135,24 @@ function(configure_openssl)
-         set(VERBOSE_OPTION OUTPUT_QUIET)
-     endif()
- 
-+    if (CMAKE_SYSTEM_NAME MATCHES "Emscripten")
-+        set(EMSCRIPTEN_CMAKE_WRAPPER "emcmake")
-+        find_program(EMCC emcc REQUIRED)
-+        set(EMSCRIPTEN_LINKER ${EMCC})
-+        list(APPEND CONFIGURE_COMMAND wasm64)
-+    else()
-+        set(EMSCRIPTEN_CMAKE_WRAPPER "")
-+        set(EMSCRIPTEN_LINKER ${CMAKE_LINKER})
-+    endif ()
-+
-     execute_process(
--        COMMAND ${CMAKE_COMMAND} -E env
-+        COMMAND ${EMSCRIPTEN_CMAKE_WRAPPER} ${CMAKE_COMMAND} -E env
-             "CFLAGS=${CMAKE_C_FLAGS}"
-             "CXXFLAGS=${CMAKE_CXX_FLAGS}"
-+            "LDFLAGS=${CMAKE_CXX_LINK_FLAGS}"
-+            "CC=${CMAKE_C_COMPILER}"
-+            "CXX=${CMAKE_CXX_COMPILER}"
-+            "LD=${EMSCRIPTEN_LINKER}"
-             ${CONFIGURE_COMMAND}
-         WORKING_DIRECTORY ${CONFIGURE_BUILD_DIR}
-         ${VERBOSE_OPTION}

cpmfile.json

@@ -202,8 +202,7 @@
             "0001-cpmutil-compat.patch",
             "0002-use-ccache.patch",
             "0003-use-cmake-compiler-flags.patch",
-            "0004-use-shell-wrapper.patch",
-            "0005-cross-compile.patch"
+            "0004-use-shell-wrapper.patch"
         ],
         "repo": "jimmy-park/openssl-cmake",
         "version": "3.6.2"
@@ -334,11 +333,6 @@
         "repo": "herumi/xbyak",
         "version": "v7.40.1"
     },
-    "zbic": {
-        "hash": "fbe2f37986377d7f0d96ae3224c80b5971df6e8b6961f68061f1975ebb2e8cb78f07b90f014b007be4023dbf5513581504e7f7d61a5237cb2a8a7a61ae11e482",
-        "repo": "kinnay/zbic",
-        "version": "11b08f2712264bbed731545085cbd9702096ceb7"
-    },
     "zlib": {
         "hash": "16fea4df307a68cf0035858abe2fd550250618a97590e202037acd18a666f57afc10f8836cbbd472d54a0e76539d0e558cb26f059d53de52ff90634bbf4f47d4",
         "min_version": "1.2",

docs/CrossCompile.md

@@ -6,31 +6,10 @@ General guide for cross compiling.
 
 A painless guide for cross compilation (or to test NCE) from a x86_64 system without polluting your main.
 
-```sh
-#!/bin/sh
-ISO_FILE="debian-13.0.0-arm64-netinst.iso"
-DISK_FILE="debian-13-arm64-ci.qcow2"
-which -- qemu-system-aarch64 || sudo pkg install qemu
-[ -f "$ISO_FILE" ] || wget https://cdimage.debian.org/debian-cd/current/arm64/iso-cd/$ISO_FILE
-[ -f "$DISK_FILE" ] || qemu-img create -f qcow2 "$DISK_FILE" 30G
-QEMU_FLAGS=(
-    -M virt
-    -m 2G
-    -cpu max
-    -bios /usr/local/share/qemu/edk2-aarch64-code.fd
-    -drive if=none,file=$ISO_FILE,format=raw,id=cdrom
-    -drive if=none,file=$DISK_FILE,id=hd0,format=qcow2
-    -device scsi-cd,drive=cdrom
-    -device virtio-blk-device,drive=hd0
-    -device virtio-gpu-pci
-    -device usb-ehci
-    -device usb-kbd
-    -device intel-hda
-    -device hda-output
-    -nic user,model=virtio-net-pci
-)
-qemu-system-aarch64 "${QEMU_FLAGS[@]}"
-```
+- Install QEMU: `sudo pkg install qemu`
+- Download Debian 13: `wget https://cdimage.debian.org/debian-cd/current/arm64/iso-cd/debian-13.0.0-arm64-netinst.iso`
+- Create a system disk: `qemu-img create -f qcow2 debian-13-arm64-ci.qcow2 30G`
+- Run the VM: `qemu-system-aarch64 -M virt -m 2G -cpu max -bios /usr/local/share/qemu/edk2-aarch64-code.fd -drive if=none,file=debian-13.0.0-arm64-netinst.iso,format=raw,id=cdrom -device scsi-cd,drive=cdrom -drive if=none,file=debian-13-arm64-ci.qcow2,id=hd0,format=qcow2 -device virtio-blk-device,drive=hd0 -device virtio-gpu-pci -device usb-ehci -device usb-kbd -device intel-hda -device hda-output -nic user,model=virtio-net-pci`
 
 ## Gentoo
 

externals/CMakeLists.txt

@@ -48,11 +48,6 @@ if (NOT TARGET stb::headers)
     add_library(stb::headers ALIAS stb)
 endif()
 
-AddJsonPackage(NAME zbic DOWNLOAD_ONLY)
-set(ZBIC_INCLUDE_DIR
-    "${zbic_SOURCE_DIR}/src"
-    PARENT_SCOPE)
-
 # ItaniumDemangle (Windows only)
 if (WIN32 AND NOT TARGET LLVM::Demangle)
     add_library(demangle demangle/ItaniumDemangle.cpp)

externals/cmake-modules/DetectArchitecture.cmake

@@ -101,11 +101,6 @@ endfunction()
 # ia64 exists
 # the rest exist, but are probably less popular than ia64
 
-detect_architecture_symbols(
-    ARCH wasm
-    SYMBOLS
-        "__EMSCRIPTEN__")
-
 detect_architecture_symbols(
     ARCH arm64
     SYMBOLS
@@ -208,6 +203,11 @@ detect_architecture_symbols(
         "__loongarch__"
         "__loongarch64")
 
+detect_architecture_symbols(
+    ARCH wasm
+    SYMBOLS
+        "__EMSCRIPTEN__")
+
 # "generic" target
 # If you have reached this point, you're on some as-of-yet unsupported architecture.
 # See the docs up above for known unsupported architectures

src/common/CMakeLists.txt

@@ -137,8 +137,6 @@ add_library(
     uuid.cpp
     uuid.h
     vector_math.h
-    zbic_compression.cpp
-    zbic_compression.h
     zstd_compression.cpp
     zstd_compression.h
     fs/ryujinx_compat.h fs/ryujinx_compat.cpp
@@ -149,10 +147,6 @@ add_library(
     net/net.h net/net.cpp
     container/unordered_map.h container/unordered_set.h)
 
-set_source_files_properties(zbic_compression.cpp PROPERTIES
-    INCLUDE_DIRECTORIES "${ZBIC_INCLUDE_DIR}"
-    COMPILE_OPTIONS "$<$<CXX_COMPILER_ID:Clang,GNU>:-Wno-unused-function;-Wno-missing-declarations;-Wno-shadow>")
-
 if(WIN32)
     target_sources(common PRIVATE windows/timer_resolution.cpp
         windows/timer_resolution.h)

src/common/zbic_compression.cpp

@@ -1,46 +0,0 @@
-// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
-// SPDX-License-Identifier: GPL-3.0-or-later
-
-#include <cstring>
-
-#include "common/zbic_compression.h"
-
-#define ZSTD_ZBIC_SUPPORT 1
-#define ZSTDLIB_VISIBLE static
-#define ZSTDLIB_HIDDEN static
-#define ZSTDERRORLIB_VISIBLE static
-#define ZSTDERRORLIB_HIDDEN static
-#undef ZSTD_MULTITHREAD
-
-#if defined(__ANDROID__)
-#undef _GNU_SOURCE
-#endif
-
-#include "zstd.h"
-#define g_ZSTD_threading_useless_symbol g_ZSTD_zbic_threading_useless_symbol
-#include "zstd.c"
-#undef g_ZSTD_threading_useless_symbol
-
-namespace Common::Compression {
-
-bool IsZBIC(std::span<const u8> src) {
-    if (src.size() < sizeof(u32)) {
-        return false;
-    }
-    u32 magic = 0;
-    std::memcpy(&magic, src.data(), sizeof(u32));
-    return magic == ZSTD_MAGICNUMBER; // 0x4349425A ("ZBIC")
-}
-
-int DecompressDataZBIC(std::span<u8> dst, std::span<const u8> src) {
-    if (dst.empty() || src.empty()) {
-        return -1;
-    }
-    const size_t res = ZSTD_decompress(dst.data(), dst.size(), src.data(), src.size());
-    if (ZSTD_isError(res)) {
-        return -1;
-    }
-    return static_cast<int>(res);
-}
-
-} // namespace Common::Compression

src/common/zbic_compression.h

@@ -1,15 +0,0 @@
-// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
-// SPDX-License-Identifier: GPL-3.0-or-later
-
-#pragma once
-
-#include <span>
-#include "common/common_types.h"
-
-namespace Common::Compression {
-
-[[nodiscard]] bool IsZBIC(std::span<const u8> src);
-
-[[nodiscard]] int DecompressDataZBIC(std::span<u8> dst, std::span<const u8> src);
-
-} // namespace Common::Compression

src/core/loader/nso.cpp

@@ -7,14 +7,12 @@
 #include <algorithm>
 #include <cinttypes>
 #include <cstring>
-#include <span>
 #include <vector>
 
 #include "common/common_funcs.h"
 #include "common/hex_util.h"
 #include "common/logging.h"
 #include "common/lz4_compression.h"
-#include "common/zbic_compression.h"
 #include "common/settings.h"
 #include "common/swap.h"
 #include "core/core.h"
@@ -106,36 +104,11 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::KProcess& process, Core::
         for (std::size_t i = 0; i < nso_header.segments.size(); ++i) {
             nso_file.Read(compressed_data.data(), nso_header.segments_compressed_size[i], nso_header.segments[i].offset);
             if (nso_header.IsSegmentCompressed(i)) {
-                if (nso_header.IsZBICCompressed()) {
-                    // ZBIC compression
-                    const int r = Common::Compression::DecompressDataZBIC(
-                        std::span<u8>{decompressed_size}.first(nso_header.segments[i].size),
-                        std::span<const u8>{compressed_data}.first(nso_header.segments_compressed_size[i])
-                    );
-                    ASSERT(r > 0);
-                } else {
-                    // LZ4 compression
-                    int r = Common::Compression::DecompressDataLZ4(
-                        decompressed_size.data(),
-                        nso_header.segments[i].size,
-                        compressed_data.data(),
-                        nso_header.segments_compressed_size[i]
-                    );
-                    ASSERT(r == int(nso_header.segments[i].size));
-                }
-
-                std::memcpy(
-                    codeset.memory.data() + module_start + nso_header.segments[i].location,
-                    decompressed_size.data(),
-                    nso_header.segments[i].size
-                );
+                int r = Common::Compression::DecompressDataLZ4(decompressed_size.data(), nso_header.segments[i].size, compressed_data.data(), nso_header.segments_compressed_size[i]);
+                ASSERT(r == int(nso_header.segments[i].size));
+                std::memcpy(codeset.memory.data() + module_start + nso_header.segments[i].location, decompressed_size.data(), nso_header.segments[i].size);
             } else {
-                // Not compressed
-                std::memcpy(
-                    codeset.memory.data() + module_start + nso_header.segments[i].location,
-                    compressed_data.data(),
-                    nso_header.segments[i].size
-                );
+                std::memcpy(codeset.memory.data() + module_start + nso_header.segments[i].location, compressed_data.data(), nso_header.segments[i].size);
             }
             codeset.segments[i].addr = module_start + nso_header.segments[i].location;
             codeset.segments[i].offset = module_start + nso_header.segments[i].location;

src/core/loader/nso.h

@@ -1,6 +1,3 @@
-// SPDX-FileCopyrightText: Copyright 2026 Eden Emulator Project
-// SPDX-License-Identifier: GPL-3.0-or-later
-
 // SPDX-FileCopyrightText: Copyright 2018 yuzu Emulator Project
 // SPDX-License-Identifier: GPL-2.0-or-later
 
@@ -61,9 +58,6 @@ struct NSOHeader {
     std::array<SHA256Hash, 3> segment_hashes;
 
     bool IsSegmentCompressed(size_t segment_num) const;
-    bool IsZBICCompressed() const {
-        return ((flags >> 7) & 1) != 0;
-    }
 };
 static_assert(sizeof(NSOHeader) == 0x100, "NSOHeader has incorrect size.");
 static_assert(std::is_trivially_copyable_v<NSOHeader>, "NSOHeader must be trivially copyable.");

src/core/memory.cpp

@@ -757,7 +757,7 @@ struct Memory::Impl {
         };
         auto& gpu = system.GPU();
         gpu_device_memory->ApplyOpOnPointer(
-            p, scratch_buffers[core], [&](DAddr address) { gpu.InvalidateRegion(address, size); });
+            p, scratch_buffers[core], [&](DAddr address) { gpu.InvalidateRegion(address, size, true); });
     }
 
     Core::System& system;

src/dynarmic/src/dynarmic/backend/riscv64/emit_riscv64_a32.cpp

@@ -197,15 +197,15 @@ void EmitA32Terminal(biscuit::Assembler& as, EmitContext& ctx, IR::Term::CheckHa
 
 void EmitA32LeafTerminal(biscuit::Assembler& as, EmitContext& ctx, IR::Term::LeafTerminal terminal, IR::LocationDescriptor initial_location, bool is_single_step) {
     if (auto const x = std::get_if<IR::Term::ReturnToDispatch>(&terminal))
-        return EmitA32Terminal(as, ctx, *x, initial_location, is_single_step);
+        return EmitA32LeafTerminal(as, ctx, *x, initial_location, is_single_step);
     if (auto const x = std::get_if<IR::Term::LinkBlock>(&terminal))
-        return EmitA32Terminal(as, ctx, *x, initial_location, is_single_step);
+        return EmitA32LeafTerminal(as, ctx, *x, initial_location, is_single_step);
     if (auto const x = std::get_if<IR::Term::LinkBlockFast>(&terminal))
-        return EmitA32Terminal(as, ctx, *x, initial_location, is_single_step);
+        return EmitA32LeafTerminal(as, ctx, *x, initial_location, is_single_step);
     if (auto const x = std::get_if<IR::Term::PopRSBHint>(&terminal))
-        return EmitA32Terminal(as, ctx, *x, initial_location, is_single_step);
+        return EmitA32LeafTerminal(as, ctx, *x, initial_location, is_single_step);
     if (auto const x = std::get_if<IR::Term::FastDispatchHint>(&terminal))
-        return EmitA32Terminal(as, ctx, *x, initial_location, is_single_step);
+        return EmitA32LeafTerminal(as, ctx, *x, initial_location, is_single_step);
     UNREACHABLE();
 }
 

src/video_core/gpu.cpp

@@ -226,7 +226,17 @@ struct GPU::Impl {
     }
 
     /// Notify rasterizer that any caches of the specified region should be invalidated
-    void InvalidateRegion(DAddr addr, u64 size) {
+    void InvalidateRegion(DAddr addr, u64 size, bool preserve_gpu_writes) {
+        VideoCore::RasterizerInterface* rasterizer = renderer->ReadRasterizer();
+        if (preserve_gpu_writes && rasterizer->MustFlushRegion(addr, size, VideoCommon::CacheType::BufferCache)) {
+            const u64 fence = RequestSyncOperation([rasterizer, addr, size] {
+                rasterizer->FlushRegion(addr, size, VideoCommon::CacheType::BufferCache);
+                rasterizer->OnCacheInvalidation(addr, size);
+            });
+            gpu_thread.TickGPU(is_async);
+            WaitForSyncOperation(fence);
+            return;
+        }
         gpu_thread.InvalidateRegion(addr, size);
     }
 
@@ -518,8 +528,8 @@ void GPU::FlushRegion(DAddr addr, u64 size) {
     impl->FlushRegion(addr, size);
 }
 
-void GPU::InvalidateRegion(DAddr addr, u64 size) {
-    impl->InvalidateRegion(addr, size);
+void GPU::InvalidateRegion(DAddr addr, u64 size, bool preserve_gpu_writes) {
+    impl->InvalidateRegion(addr, size, preserve_gpu_writes);
 }
 
 bool GPU::OnCPUWrite(DAddr addr, u64 size) {

src/video_core/gpu.h

@@ -248,7 +248,7 @@ public:
     void FlushRegion(DAddr addr, u64 size);
 
     /// Notify rasterizer that any caches of the specified region should be invalidated
-    void InvalidateRegion(DAddr addr, u64 size);
+    void InvalidateRegion(DAddr addr, u64 size, bool preserve_gpu_writes = false);
 
     /// Notify rasterizer that CPU is trying to write this area. It returns true if the area is
     /// sensible, false otherwise, addr and size must be a valid combination