"Merge Tagged PR 1340"

Improvements to GPU synchronization & various refactoring

.gitmodules

@@ -1,6 +1,6 @@
 [submodule "inih"]
     path = externals/inih/inih
-    url = https://github.com/svn2github/inih
+    url = https://github.com/benhoyt/inih.git
 [submodule "cubeb"]
     path = externals/cubeb
     url = https://github.com/kinetiknz/cubeb.git

CMakeLists.txt

@@ -161,7 +161,7 @@ macro(yuzu_find_packages)
     #    Cmake Pkg Prefix  Version     Conan Pkg
         "Boost             1.73        boost/1.73.0"
         "Catch2            2.13        catch2/2.13.0"
-        "fmt               7.0         fmt/7.0.3"
+        "fmt               7.1         fmt/7.1.0"
     # can't use until https://github.com/bincrafters/community/issues/1173
         #"libzip            1.5         libzip/1.5.2@bincrafters/stable"
         "lz4               1.8         lz4/1.9.2"
@@ -263,6 +263,7 @@ if (CONAN_REQUIRED_LIBS)
         libzip:with_openssl=False
         libzip:enable_windows_crypto=False
     )
+
     conan_check(VERSION 1.24.0 REQUIRED)
     # Add the bincrafters remote
     conan_add_remote(NAME bincrafters
@@ -354,6 +355,19 @@ if (NOT LIBUSB_FOUND)
     set(LIBUSB_LIBRARIES usb)
 endif()
 
+# Use system installed ffmpeg.
+if (NOT MSVC)
+    find_package(FFmpeg REQUIRED)
+else()
+    set(FFMPEG_EXT_NAME "ffmpeg-4.2.1")
+    set(FFMPEG_PATH "${CMAKE_BINARY_DIR}/externals/${FFMPEG_EXT_NAME}")
+    download_bundled_external("ffmpeg/" ${FFMPEG_EXT_NAME} "")
+    set(FFMPEG_FOUND YES)
+    set(FFMPEG_INCLUDE_DIR "${FFMPEG_PATH}/include" CACHE PATH "Path to FFmpeg headers" FORCE)
+    set(FFMPEG_LIBRARY_DIR "${FFMPEG_PATH}/bin" CACHE PATH "Path to FFmpeg library" FORCE)
+    set(FFMPEG_DLL_DIR "${FFMPEG_PATH}/bin" CACHE PATH "Path to FFmpeg dll's" FORCE)
+endif()
+
 # Prefer the -pthread flag on Linux.
 set(THREADS_PREFER_PTHREAD_FLAG ON)
 find_package(Threads REQUIRED)

CMakeModules/CopyYuzuFFmpegDeps.cmake

@@ -0,0 +1,10 @@
+function(copy_yuzu_FFmpeg_deps target_dir)
+    include(WindowsCopyFiles)
+    set(DLL_DEST "${CMAKE_BINARY_DIR}/bin/$<CONFIG>/")
+    windows_copy_files(${target_dir} ${FFMPEG_DLL_DIR} ${DLL_DEST}
+        avcodec-58.dll
+        avutil-56.dll
+        swresample-3.dll
+        swscale-5.dll
+    )
+endfunction(copy_yuzu_FFmpeg_deps)

externals/CMakeLists.txt

@@ -73,23 +73,29 @@ if (NOT LIBZIP_FOUND)
 endif()
 
 if (ENABLE_WEB_SERVICE)
-    # LibreSSL
-    set(LIBRESSL_SKIP_INSTALL ON CACHE BOOL "")
-    add_subdirectory(libressl EXCLUDE_FROM_ALL)
-    target_include_directories(ssl INTERFACE ./libressl/include)
-    target_compile_definitions(ssl PRIVATE -DHAVE_INET_NTOP)
-    get_directory_property(OPENSSL_LIBRARIES
-        DIRECTORY libressl
-        DEFINITION OPENSSL_LIBS)
-
-    # lurlparser
-    add_subdirectory(lurlparser EXCLUDE_FROM_ALL)
+    find_package(OpenSSL 1.1)
+    if (OPENSSL_FOUND)
+        set(OPENSSL_LIBRARIES OpenSSL::SSL OpenSSL::Crypto)
+    else()
+        # LibreSSL
+        set(LIBRESSL_SKIP_INSTALL ON CACHE BOOL "")
+        set(OPENSSLDIR "/etc/ssl/")
+        add_subdirectory(libressl EXCLUDE_FROM_ALL)
+        target_include_directories(ssl INTERFACE ./libressl/include)
+        target_compile_definitions(ssl PRIVATE -DHAVE_INET_NTOP)
+        get_directory_property(OPENSSL_LIBRARIES
+            DIRECTORY libressl
+            DEFINITION OPENSSL_LIBS)
+    endif()
 
     # httplib
     add_library(httplib INTERFACE)
     target_include_directories(httplib INTERFACE ./httplib)
     target_compile_definitions(httplib INTERFACE -DCPPHTTPLIB_OPENSSL_SUPPORT)
     target_link_libraries(httplib INTERFACE ${OPENSSL_LIBRARIES})
+    if (WIN32)
+        target_link_libraries(httplib INTERFACE crypt32 cryptui ws2_32)
+    endif()
 endif()
 
 # Opus

externals/find-modules/FindFFmpeg.cmake

@@ -0,0 +1,100 @@
+# - Try to find ffmpeg libraries (libavcodec, libavformat and libavutil)
+# Once done this will define
+#
+# FFMPEG_FOUND - system has ffmpeg or libav
+# FFMPEG_INCLUDE_DIR - the ffmpeg include directory
+# FFMPEG_LIBRARIES - Link these to use ffmpeg
+# FFMPEG_LIBAVCODEC
+# FFMPEG_LIBAVFORMAT
+# FFMPEG_LIBAVUTIL
+#
+# Copyright (c) 2008 Andreas Schneider <[email protected]>
+# Modified for other libraries by Lasse Kärkkäinen <tronic>
+# Modified for Hedgewars by Stepik777
+# Modified for FFmpeg-example Tuukka Pasanen 2018
+# Modified for yuzu toastUnlimted 2020
+#
+# Redistribution and use is allowed according to the terms of the New
+# BSD license.
+#
+
+include(FindPackageHandleStandardArgs)
+
+find_package_handle_standard_args(FFMPEG
+  FOUND_VAR FFMPEG_FOUND
+  REQUIRED_VARS
+      FFMPEG_LIBRARY
+      FFMPEG_INCLUDE_DIR
+  VERSION_VAR FFMPEG_VERSION
+)
+
+if(FFMPEG_LIBRARIES AND FFMPEG_INCLUDE_DIR)
+  # in cache already
+  set(FFMPEG_FOUND TRUE)
+else()
+  # use pkg-config to get the directories and then use these values
+  # in the FIND_PATH() and FIND_LIBRARY() calls
+  find_package(PkgConfig)
+  if(PKG_CONFIG_FOUND)
+    pkg_check_modules(_FFMPEG_AVCODEC libavcodec)
+    pkg_check_modules(_FFMPEG_AVUTIL libavutil)
+    pkg_check_modules(_FFMPEG_SWSCALE libswscale)
+  endif()
+
+  find_path(FFMPEG_AVCODEC_INCLUDE_DIR
+    NAMES libavcodec/avcodec.h
+    PATHS ${_FFMPEG_AVCODEC_INCLUDE_DIRS}
+      /usr/include
+      /usr/local/include
+      /opt/local/include
+      /sw/include
+    PATH_SUFFIXES ffmpeg libav)
+
+  find_library(FFMPEG_LIBAVCODEC
+    NAMES avcodec
+    PATHS ${_FFMPEG_AVCODEC_LIBRARY_DIRS}
+      /usr/lib
+      /usr/local/lib
+      /opt/local/lib
+      /sw/lib)
+
+  find_library(FFMPEG_LIBAVUTIL
+    NAMES avutil
+    PATHS ${_FFMPEG_AVUTIL_LIBRARY_DIRS}
+      /usr/lib
+      /usr/local/lib
+      /opt/local/lib
+      /sw/lib)
+
+  find_library(FFMPEG_LIBSWSCALE
+    NAMES swscale
+    PATHS ${_FFMPEG_SWSCALE_LIBRARY_DIRS}
+      /usr/lib
+      /usr/local/lib
+      /opt/local/lib
+      /sw/lib)
+
+  if(FFMPEG_LIBAVCODEC AND FFMPEG_LIBAVUTIL AND FFMPEG_LIBSWSCALE)
+    set(FFMPEG_FOUND TRUE)
+  endif()
+
+  if(FFMPEG_FOUND)
+    set(FFMPEG_INCLUDE_DIR ${FFMPEG_AVCODEC_INCLUDE_DIR})
+    set(FFMPEG_LIBRARIES
+      ${FFMPEG_LIBAVCODEC}
+      ${FFMPEG_LIBAVUTIL}
+      ${FFMPEG_LIBSWSCALE})
+  endif()
+
+  if(FFMPEG_FOUND)
+    if(NOT FFMPEG_FIND_QUIETLY)
+      message(STATUS
+      "Found FFMPEG or Libav: ${FFMPEG_LIBRARIES}, ${FFMPEG_INCLUDE_DIR}")
+    endif()
+  else()
+    if(FFMPEG_FIND_REQUIRED)
+      message(FATAL_ERROR
+      "Could not find libavcodec or libavutil or libswscale")
+    endif()
+  endif()
+endif()

externals/httplib/README.md

@@ -1,4 +1,4 @@
-From https://github.com/yhirose/cpp-httplib/tree/fce8e6fefdab4ad48bc5b25c98e5ebfda4f3cf53
+From https://github.com/yhirose/cpp-httplib/tree/ff5677ad197947177c158fe857caff4f0e242045 with https://github.com/yhirose/cpp-httplib/pull/701
 
 MIT License
 

externals/lurlparser/CMakeLists.txt

@@ -1,8 +0,0 @@
-add_library(lurlparser
-        LUrlParser.cpp
-        LUrlParser.h
-)
-
-create_target_directory_groups(lurlparser)
-
-target_include_directories(lurlparser INTERFACE ${CMAKE_CURRENT_SOURCE_DIR})

externals/lurlparser/LUrlParser.cpp

@@ -1,265 +0,0 @@
-/*
- * Lightweight URL & URI parser (RFC 1738, RFC 3986)
- * https://github.com/corporateshark/LUrlParser
- *
- * The MIT License (MIT)
- *
- * Copyright (C) 2015 Sergey Kosarevsky (sk@linderdaum.com)
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to deal
- * in the Software without restriction, including without limitation the rights
- * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
- * copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#include "LUrlParser.h"
-
-#include <algorithm>
-#include <cstring>
-#include <stdlib.h>
-
-// check if the scheme name is valid
-static bool IsSchemeValid( const std::string& SchemeName )
-{
-    for ( auto c : SchemeName  )
-    {
-        if ( !isalpha( c ) && c != '+' && c != '-' && c != '.' ) return false;
-    }
-
-    return true;
-}
-
-bool LUrlParser::clParseURL::GetPort( int* OutPort ) const
-{
-    if ( !IsValid() ) { return false; }
-
-    int Port = atoi( m_Port.c_str() );
-
-    if ( Port <= 0 || Port > 65535 ) { return false; }
-
-    if ( OutPort ) { *OutPort = Port; }
-
-    return true;
-}
-
-// based on RFC 1738 and RFC 3986
-LUrlParser::clParseURL LUrlParser::clParseURL::ParseURL( const std::string& URL )
-{
-    LUrlParser::clParseURL Result;
-
-    const char* CurrentString = URL.c_str();
-
-    /*
-     *	<scheme>:<scheme-specific-part>
-     *	<scheme> := [a-z\+\-\.]+
-     *	For resiliency, programs interpreting URLs should treat upper case letters as equivalent to lower case in scheme names
-     */
-
-    // try to read scheme
-    {
-        const char* LocalString = strchr( CurrentString, ':' );
-
-        if ( !LocalString )
-        {
-            return clParseURL( LUrlParserError_NoUrlCharacter );
-        }
-
-        // save the scheme name
-        Result.m_Scheme = std::string( CurrentString, LocalString - CurrentString );
-
-        if ( !IsSchemeValid( Result.m_Scheme ) )
-        {
-            return clParseURL( LUrlParserError_InvalidSchemeName );
-        }
-
-        // scheme should be lowercase
-        std::transform( Result.m_Scheme.begin(), Result.m_Scheme.end(), Result.m_Scheme.begin(), ::tolower );
-
-        // skip ':'
-        CurrentString = LocalString+1;
-    }
-
-    /*
-     *	//<user>:<password>@<host>:<port>/<url-path>
-     *	any ":", "@" and "/" must be normalized
-     */
-
-    // skip "//"
-    if ( *CurrentString++ != '/' ) return clParseURL( LUrlParserError_NoDoubleSlash );
-    if ( *CurrentString++ != '/' ) return clParseURL( LUrlParserError_NoDoubleSlash );
-
-    // check if the user name and password are specified
-    bool bHasUserName = false;
-
-    const char* LocalString = CurrentString;
-
-    while ( *LocalString )
-    {
-        if ( *LocalString == '@' )
-        {
-            // user name and password are specified
-            bHasUserName = true;
-            break;
-        }
-        else if ( *LocalString == '/' )
-        {
-            // end of <host>:<port> specification
-            bHasUserName = false;
-            break;
-        }
-
-        LocalString++;
-    }
-
-    // user name and password
-    LocalString = CurrentString;
-
-    if ( bHasUserName )
-    {
-        // read user name
-        while ( *LocalString && *LocalString != ':' && *LocalString != '@' ) LocalString++;
-
-        Result.m_UserName = std::string( CurrentString, LocalString - CurrentString );
-
-        // proceed with the current pointer
-        CurrentString = LocalString;
-
-        if ( *CurrentString == ':' )
-        {
-            // skip ':'
-            CurrentString++;
-
-            // read password
-            LocalString = CurrentString;
-
-            while ( *LocalString && *LocalString != '@' ) LocalString++;
-
-            Result.m_Password = std::string( CurrentString, LocalString - CurrentString );
-
-            CurrentString = LocalString;
-        }
-
-        // skip '@'
-        if ( *CurrentString != '@' )
-        {
-            return clParseURL( LUrlParserError_NoAtSign );
-        }
-
-        CurrentString++;
-    }
-
-    bool bHasBracket = ( *CurrentString == '[' );
-
-    // go ahead, read the host name
-    LocalString = CurrentString;
-
-    while ( *LocalString )
-    {
-        if ( bHasBracket && *LocalString == ']' )
-        {
-            // end of IPv6 address
-            LocalString++;
-            break;
-        }
-        else if ( !bHasBracket && ( *LocalString == ':' || *LocalString == '/' ) )
-        {
-            // port number is specified
-            break;
-        }
-
-        LocalString++;
-    }
-
-    Result.m_Host = std::string( CurrentString, LocalString - CurrentString );
-
-    CurrentString = LocalString;
-
-    // is port number specified?
-    if ( *CurrentString == ':' )
-    {
-        CurrentString++;
-
-        // read port number
-        LocalString = CurrentString;
-
-        while ( *LocalString && *LocalString != '/' ) LocalString++;
-
-        Result.m_Port = std::string( CurrentString, LocalString - CurrentString );
-
-        CurrentString = LocalString;
-    }
-
-    // end of string
-    if ( !*CurrentString )
-    {
-        Result.m_ErrorCode = LUrlParserError_Ok;
-
-        return Result;
-    }
-
-    // skip '/'
-    if ( *CurrentString != '/' )
-    {
-        return clParseURL( LUrlParserError_NoSlash );
-    }
-
-    CurrentString++;
-
-    // parse the path
-    LocalString = CurrentString;
-
-    while ( *LocalString && *LocalString != '#' && *LocalString != '?' ) LocalString++;
-
-    Result.m_Path = std::string( CurrentString, LocalString - CurrentString );
-
-    CurrentString = LocalString;
-
-    // check for query
-    if ( *CurrentString == '?' )
-    {
-        // skip '?'
-        CurrentString++;
-
-        // read query
-        LocalString = CurrentString;
-
-        while ( *LocalString && *LocalString != '#' ) LocalString++;
-
-        Result.m_Query = std::string( CurrentString, LocalString - CurrentString );
-
-        CurrentString = LocalString;
-    }
-
-    // check for fragment
-    if ( *CurrentString == '#' )
-    {
-        // skip '#'
-        CurrentString++;
-
-        // read fragment
-        LocalString = CurrentString;
-
-        while ( *LocalString ) LocalString++;
-
-        Result.m_Fragment = std::string( CurrentString, LocalString - CurrentString );
-
-        CurrentString = LocalString;
-    }
-
-    Result.m_ErrorCode = LUrlParserError_Ok;
-
-    return Result;
-}

externals/lurlparser/LUrlParser.h

@@ -1,78 +0,0 @@
-/*
- * Lightweight URL & URI parser (RFC 1738, RFC 3986)
- * https://github.com/corporateshark/LUrlParser
- *
- * The MIT License (MIT)
- *
- * Copyright (C) 2015 Sergey Kosarevsky (sk@linderdaum.com)
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to deal
- * in the Software without restriction, including without limitation the rights
- * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
- * copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#pragma once
-
-#include <string>
-
-namespace LUrlParser
-{
-enum LUrlParserError
-{
-    LUrlParserError_Ok = 0,
-    LUrlParserError_Uninitialized = 1,
-    LUrlParserError_NoUrlCharacter = 2,
-    LUrlParserError_InvalidSchemeName = 3,
-    LUrlParserError_NoDoubleSlash = 4,
-    LUrlParserError_NoAtSign = 5,
-    LUrlParserError_UnexpectedEndOfLine = 6,
-    LUrlParserError_NoSlash = 7,
-};
-
-class clParseURL
-{
-public:
-    LUrlParserError m_ErrorCode;
-    std::string m_Scheme;
-    std::string m_Host;
-    std::string m_Port;
-    std::string m_Path;
-    std::string m_Query;
-    std::string m_Fragment;
-    std::string m_UserName;
-    std::string m_Password;
-
-    clParseURL()
-            : m_ErrorCode( LUrlParserError_Uninitialized )
-    {}
-
-    /// return 'true' if the parsing was successful
-    bool IsValid() const { return m_ErrorCode == LUrlParserError_Ok; }
-
-    /// helper to convert the port number to int, return 'true' if the port is valid (within the 0..65535 range)
-    bool GetPort( int* OutPort ) const;
-
-    /// parse the URL
-    static clParseURL ParseURL( const std::string& URL );
-
-private:
-    explicit clParseURL( LUrlParserError ErrorCode )
-            : m_ErrorCode( ErrorCode )
-    {}
-};
-
-} // namespace LUrlParser

externals/lurlparser/README.md

@@ -1,19 +0,0 @@
-From https://github.com/corporateshark/LUrlParser/commit/455d5e2d27e3946f11ad0328fee9ee2628e6a8e2
-
-MIT License
-
-===
-
-Lightweight URL & URI parser (RFC 1738, RFC 3986)
-
-(C) Sergey Kosarevsky, 2015
-
-@corporateshark sk@linderdaum.com
-
-http://www.linderdaum.com
-
-http://blog.linderdaum.com
-
-=============================
-
-A tiny and lightweight URL & URI parser (RFC 1738, RFC 3986) written in C++.

externals/microprofile/microprofile.h

@@ -902,8 +902,10 @@ inline uint16_t MicroProfileGetGroupIndex(MicroProfileToken t)
 #include <windows.h>
 #define snprintf _snprintf
 
+#ifdef _MSC_VER
 #pragma warning(push)
 #pragma warning(disable: 4244)
+#endif
 int64_t MicroProfileTicksPerSecondCpu()
 {
     static int64_t nTicksPerSecond = 0;
@@ -946,7 +948,11 @@ typedef HANDLE MicroProfileThread;
 DWORD _stdcall ThreadTrampoline(void* pFunc)
 {
     MicroProfileThreadFunc F = (MicroProfileThreadFunc)pFunc;
-    return (uint32_t)F(0);
+
+    // The return value of F will always return a void*, however, this is for
+    // compatibility with pthreads. The underlying "address" of the pointer
+    // is always a 32-bit value, so this cast is safe to perform.
+    return static_cast<DWORD>(reinterpret_cast<uint64_t>(F(0)));
 }
 
 inline void MicroProfileThreadStart(MicroProfileThread* pThread, MicroProfileThreadFunc Func)
@@ -1742,10 +1748,10 @@ void MicroProfileFlip()
                             }
                         }
                     }
-                    for(uint32_t i = 0; i < MICROPROFILE_MAX_GROUPS; ++i)
+                    for(uint32_t j = 0; j < MICROPROFILE_MAX_GROUPS; ++j)
                     {
-                        pLog->nGroupTicks[i] += nGroupTicks[i];
-                        pFrameGroup[i] += nGroupTicks[i];
+                        pLog->nGroupTicks[j] += nGroupTicks[j];
+                        pFrameGroup[j] += nGroupTicks[j];
                     }
                     pLog->nStackPos = nStackPos;
                 }
@@ -3328,7 +3334,7 @@ bool MicroProfileIsLocalThread(uint32_t nThreadId)
 #endif
 #else
 
-bool MicroProfileIsLocalThread(uint32_t nThreadId){return false;}
+bool MicroProfileIsLocalThread([[maybe_unused]] uint32_t nThreadId) { return false; }
 void MicroProfileStopContextSwitchTrace(){}
 void MicroProfileStartContextSwitchTrace(){}
 
@@ -3576,7 +3582,7 @@ int MicroProfileGetGpuTickReference(int64_t* pOutCpu, int64_t* pOutGpu)
 
 #undef S
 
-#ifdef _WIN32
+#ifdef _MSC_VER
 #pragma warning(pop)
 #endif
 

src/CMakeLists.txt

@@ -32,7 +32,6 @@ if (MSVC)
     # /Zc:inline          - Let codegen omit inline functions in object files
     # /Zc:throwingNew     - Let codegen assume `operator new` (without std::nothrow) will never return null
     add_compile_options(
-        /W3
         /MP
         /Zi
         /Zo
@@ -43,6 +42,13 @@ if (MSVC)
         /Zc:externConstexpr
         /Zc:inline
         /Zc:throwingNew
+
+        # Warnings
+        /W3
+        /we4547 # 'operator' : operator before comma has no effect; expected operator with side-effect
+        /we4549 # 'operator1': operator before comma has no effect; did you intend 'operator2'?
+        /we4555 # Expression has no effect; expected expression with side-effect
+        /we4834 # Discarding return value of function with 'nodiscard' attribute
     )
 
     # /GS- - No stack buffer overflow checks
@@ -56,6 +62,7 @@ else()
         -Werror=implicit-fallthrough
         -Werror=missing-declarations
         -Werror=reorder
+        -Werror=unused-result
         -Wextra
         -Wmissing-declarations
         -Wno-attributes

src/audio_core/CMakeLists.txt

@@ -46,11 +46,17 @@ create_target_directory_groups(audio_core)
 
 if (NOT MSVC)
     target_compile_options(audio_core PRIVATE
+        -Werror=conversion
         -Werror=ignored-qualifiers
         -Werror=implicit-fallthrough
         -Werror=reorder
         -Werror=sign-compare
         -Werror=unused-variable
+
+        $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter>
+        $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable>
+
+        -Wno-sign-conversion
     )
 endif()
 

src/audio_core/algorithm/filter.cpp

@@ -55,7 +55,8 @@ void Filter::Process(std::vector<s16>& signal) {
 /// @param total_count The total number of biquads to be cascaded.
 /// @param index 0-index of the biquad to calculate the Q value for.
 static double CascadingBiquadQ(std::size_t total_count, std::size_t index) {
-    const double pole = M_PI * (2 * index + 1) / (4.0 * total_count);
+    const auto pole =
+        M_PI * static_cast<double>(2 * index + 1) / (4.0 * static_cast<double>(total_count));
     return 1.0 / (2.0 * std::cos(pole));
 }
 

src/audio_core/algorithm/interpolate.cpp

@@ -146,7 +146,7 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input,
         return {};
 
     if (ratio <= 0) {
-        LOG_CRITICAL(Audio, "Nonsensical interpolation ratio {}", ratio);
+        LOG_ERROR(Audio, "Nonsensical interpolation ratio {}", ratio);
         return input;
     }
 
@@ -164,7 +164,8 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input,
     const std::size_t num_frames{input.size() / 2};
 
     std::vector<s16> output;
-    output.reserve(static_cast<std::size_t>(input.size() / ratio + InterpolationState::taps));
+    output.reserve(static_cast<std::size_t>(static_cast<double>(input.size()) / ratio +
+                                            InterpolationState::taps));
 
     for (std::size_t frame{}; frame < num_frames; ++frame) {
         const std::size_t lut_index{(state.fraction >> 8) * InterpolationState::taps};

src/audio_core/command_generator.cpp

@@ -793,7 +793,6 @@ s32 CommandGenerator::DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_s
             // Decode entire frame
             if (remaining_samples >= static_cast<int>(SAMPLES_PER_FRAME)) {
                 for (std::size_t i = 0; i < SAMPLES_PER_FRAME / 2; i++) {
-
                     // Sample 1
                     const s32 s0 = SIGNED_NIBBLES[buffer[buffer_offset] >> 4];
                     const s32 s1 = SIGNED_NIBBLES[buffer[buffer_offset++] & 0xf];
@@ -802,7 +801,7 @@ s32 CommandGenerator::DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_s
                     sample_buffer[cur_mix_offset++] = sample_1;
                     sample_buffer[cur_mix_offset++] = sample_2;
                 }
-                remaining_samples -= SAMPLES_PER_FRAME;
+                remaining_samples -= static_cast<int>(SAMPLES_PER_FRAME);
                 position_in_frame += SAMPLES_PER_FRAME;
                 continue;
             }

src/audio_core/cubeb_sink.cpp

@@ -93,8 +93,10 @@ public:
                 constexpr s32 clev{707}; // center mixing level coefficient
                 constexpr s32 slev{707}; // surround mixing level coefficient
 
-                buf.push_back(left + (clev * center / 1000) + (slev * surround_left / 1000));
-                buf.push_back(right + (clev * center / 1000) + (slev * surround_right / 1000));
+                buf.push_back(static_cast<s16>(left + (clev * center / 1000) +
+                                               (slev * surround_left / 1000)));
+                buf.push_back(static_cast<s16>(right + (clev * center / 1000) +
+                                               (slev * surround_right / 1000)));
             }
             queue.Push(buf);
             return;

src/audio_core/voice_context.cpp

@@ -128,7 +128,10 @@ void ServerVoiceInfo::UpdateParameters(const VoiceInfo::InParams& voice_in,
     in_params.wave_buffer_count = voice_in.wave_buffer_count;
     in_params.wave_bufffer_head = voice_in.wave_buffer_head;
     if (behavior_info.IsFlushVoiceWaveBuffersSupported()) {
-        in_params.wave_buffer_flush_request_count += voice_in.wave_buffer_flush_request_count;
+        const auto in_request_count = in_params.wave_buffer_flush_request_count;
+        const auto voice_request_count = voice_in.wave_buffer_flush_request_count;
+        in_params.wave_buffer_flush_request_count =
+            static_cast<u8>(in_request_count + voice_request_count);
     }
     in_params.mix_id = voice_in.mix_id;
     if (behavior_info.IsSplitterSupported()) {

src/common/CMakeLists.txt

@@ -150,6 +150,8 @@ add_library(common STATIC
     scope_exit.h
     spin_lock.cpp
     spin_lock.h
+    stream.cpp
+    stream.h
     string_util.cpp
     string_util.h
     swap.h
@@ -188,8 +190,29 @@ if(ARCHITECTURE_x86_64)
     )
 endif()
 
+if (MSVC)
+  target_compile_definitions(common PRIVATE
+    # The standard library doesn't provide any replacement for codecvt yet
+    # so we can disable this deprecation warning for the time being.
+    _SILENCE_CXX17_CODECVT_HEADER_DEPRECATION_WARNING
+  )
+  target_compile_options(common PRIVATE
+    /W4
+    /WX
+  )
+else()
+  target_compile_options(common PRIVATE
+    -Werror
+  )
+endif()
+
 create_target_directory_groups(common)
 find_package(Boost 1.71 COMPONENTS context headers REQUIRED)
 
 target_link_libraries(common PUBLIC ${Boost_LIBRARIES} fmt::fmt microprofile)
-target_link_libraries(common PRIVATE lz4::lz4 zstd::zstd xbyak)
+target_link_libraries(common PRIVATE lz4::lz4 xbyak)
+if (MSVC)
+  target_link_libraries(common PRIVATE zstd::zstd)
+else()
+  target_link_libraries(common PRIVATE zstd)
+endif()

src/common/assert.h

@@ -29,22 +29,19 @@ assert_noinline_call(const Fn& fn) {
 }
 
 #define ASSERT(_a_)                                                                                \
-    do                                                                                             \
-        if (!(_a_)) {                                                                              \
-            assert_noinline_call([] { LOG_CRITICAL(Debug, "Assertion Failed!"); });                \
-        }                                                                                          \
-    while (0)
+    if (!(_a_)) {                                                                                  \
+        LOG_CRITICAL(Debug, "Assertion Failed!");                                                  \
+    }
 
 #define ASSERT_MSG(_a_, ...)                                                                       \
-    do                                                                                             \
-        if (!(_a_)) {                                                                              \
-            assert_noinline_call([&] { LOG_CRITICAL(Debug, "Assertion Failed!\n" __VA_ARGS__); }); \
-        }                                                                                          \
-    while (0)
+    if (!(_a_)) {                                                                                  \
+        LOG_CRITICAL(Debug, "Assertion Failed! " __VA_ARGS__);                                     \
+    }
 
-#define UNREACHABLE() assert_noinline_call([] { LOG_CRITICAL(Debug, "Unreachable code!"); })
+#define UNREACHABLE()                                                                              \
+    { LOG_CRITICAL(Debug, "Unreachable code!"); }
 #define UNREACHABLE_MSG(...)                                                                       \
-    assert_noinline_call([&] { LOG_CRITICAL(Debug, "Unreachable code!\n" __VA_ARGS__); })
+    { LOG_CRITICAL(Debug, "Unreachable code!\n" __VA_ARGS__); }
 
 #ifdef _DEBUG
 #define DEBUG_ASSERT(_a_) ASSERT(_a_)

src/common/fiber.cpp

@@ -79,9 +79,9 @@ void Fiber::Exit() {
     released = true;
 }
 
-void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* start_parameter) {
+void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param) {
     rewind_point = std::move(rewind_func);
-    rewind_parameter = start_parameter;
+    rewind_parameter = rewind_param;
 }
 
 void Fiber::Rewind() {
@@ -91,7 +91,7 @@ void Fiber::Rewind() {
     SwitchToFiber(impl->rewind_handle);
 }
 
-void Fiber::YieldTo(std::shared_ptr<Fiber>& from, std::shared_ptr<Fiber>& to) {
+void Fiber::YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to) {
     ASSERT_MSG(from != nullptr, "Yielding fiber is null!");
     ASSERT_MSG(to != nullptr, "Next fiber is null!");
     to->guard.lock();
@@ -161,9 +161,9 @@ Fiber::Fiber(std::function<void(void*)>&& entry_point_func, void* start_paramete
         boost::context::detail::make_fcontext(stack_base, impl->stack.size(), FiberStartFunc);
 }
 
-void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* start_parameter) {
+void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param) {
     rewind_point = std::move(rewind_func);
-    rewind_parameter = start_parameter;
+    rewind_parameter = rewind_param;
 }
 
 Fiber::Fiber() : impl{std::make_unique<FiberImpl>()} {}
@@ -199,7 +199,7 @@ void Fiber::Rewind() {
     boost::context::detail::jump_fcontext(impl->rewind_context, this);
 }
 
-void Fiber::YieldTo(std::shared_ptr<Fiber>& from, std::shared_ptr<Fiber>& to) {
+void Fiber::YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to) {
     ASSERT_MSG(from != nullptr, "Yielding fiber is null!");
     ASSERT_MSG(to != nullptr, "Next fiber is null!");
     to->guard.lock();

src/common/fiber.h

@@ -46,10 +46,10 @@ public:
 
     /// Yields control from Fiber 'from' to Fiber 'to'
     /// Fiber 'from' must be the currently running fiber.
-    static void YieldTo(std::shared_ptr<Fiber>& from, std::shared_ptr<Fiber>& to);
+    static void YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to);
     [[nodiscard]] static std::shared_ptr<Fiber> ThreadToFiber();
 
-    void SetRewindPoint(std::function<void(void*)>&& rewind_func, void* start_parameter);
+    void SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param);
 
     void Rewind();
 

src/common/file_util.cpp

@@ -472,13 +472,14 @@ u64 ScanDirectoryTree(const std::string& directory, FSTEntry& parent_entry,
 }
 
 bool DeleteDirRecursively(const std::string& directory, unsigned int recursion) {
-    const auto callback = [recursion](u64* num_entries_out, const std::string& directory,
-                                      const std::string& virtual_name) -> bool {
-        std::string new_path = directory + DIR_SEP_CHR + virtual_name;
+    const auto callback = [recursion](u64*, const std::string& directory,
+                                      const std::string& virtual_name) {
+        const std::string new_path = directory + DIR_SEP_CHR + virtual_name;
 
         if (IsDirectory(new_path)) {
-            if (recursion == 0)
+            if (recursion == 0) {
                 return false;
+            }
             return DeleteDirRecursively(new_path, recursion - 1);
         }
         return Delete(new_path);
@@ -492,7 +493,8 @@ bool DeleteDirRecursively(const std::string& directory, unsigned int recursion)
     return true;
 }
 
-void CopyDir(const std::string& source_path, const std::string& dest_path) {
+void CopyDir([[maybe_unused]] const std::string& source_path,
+             [[maybe_unused]] const std::string& dest_path) {
 #ifndef _WIN32
     if (source_path == dest_path) {
         return;
@@ -553,7 +555,7 @@ std::optional<std::string> GetCurrentDir() {
     std::string strDir = dir;
 #endif
     free(dir);
-    return std::move(strDir);
+    return strDir;
 }
 
 bool SetCurrentDir(const std::string& directory) {
@@ -772,21 +774,23 @@ std::size_t ReadFileToString(bool text_file, const std::string& filename, std::s
 
 void SplitFilename83(const std::string& filename, std::array<char, 9>& short_name,
                      std::array<char, 4>& extension) {
-    const std::string forbidden_characters = ".\"/\\[]:;=, ";
+    static constexpr std::string_view forbidden_characters = ".\"/\\[]:;=, ";
 
     // On a FAT32 partition, 8.3 names are stored as a 11 bytes array, filled with spaces.
     short_name = {{' ', ' ', ' ', ' ', ' ', ' ', ' ', ' ', '\0'}};
     extension = {{' ', ' ', ' ', '\0'}};
 
-    std::string::size_type point = filename.rfind('.');
-    if (point == filename.size() - 1)
+    auto point = filename.rfind('.');
+    if (point == filename.size() - 1) {
         point = filename.rfind('.', point);
+    }
 
     // Get short name.
     int j = 0;
     for (char letter : filename.substr(0, point)) {
-        if (forbidden_characters.find(letter, 0) != std::string::npos)
+        if (forbidden_characters.find(letter, 0) != std::string::npos) {
             continue;
+        }
         if (j == 8) {
             // TODO(Link Mauve): also do that for filenames containing a space.
             // TODO(Link Mauve): handle multiple files having the same short name.
@@ -794,14 +798,15 @@ void SplitFilename83(const std::string& filename, std::array<char, 9>& short_nam
             short_name[7] = '1';
             break;
         }
-        short_name[j++] = toupper(letter);
+        short_name[j++] = static_cast<char>(std::toupper(letter));
     }
 
     // Get extension.
     if (point != std::string::npos) {
         j = 0;
-        for (char letter : filename.substr(point + 1, 3))
-            extension[j++] = toupper(letter);
+        for (char letter : filename.substr(point + 1, 3)) {
+            extension[j++] = static_cast<char>(std::toupper(letter));
+        }
     }
 }
 

src/common/file_util.h

@@ -232,7 +232,7 @@ public:
 
     void Swap(IOFile& other) noexcept;
 
-    [[nodiscard]] bool Open(const std::string& filename, const char openmode[], int flags = 0);
+    bool Open(const std::string& filename, const char openmode[], int flags = 0);
     bool Close();
 
     template <typename T>

src/common/hex_util.h

@@ -16,14 +16,14 @@ namespace Common {
 
 [[nodiscard]] constexpr u8 ToHexNibble(char c) {
     if (c >= 65 && c <= 70) {
-        return c - 55;
+        return static_cast<u8>(c - 55);
     }
 
     if (c >= 97 && c <= 102) {
-        return c - 87;
+        return static_cast<u8>(c - 87);
     }
 
-    return c - 48;
+    return static_cast<u8>(c - 48);
 }
 
 [[nodiscard]] std::vector<u8> HexStringToVector(std::string_view str, bool little_endian);
@@ -33,11 +33,11 @@ template <std::size_t Size, bool le = false>
     std::array<u8, Size> out{};
     if constexpr (le) {
         for (std::size_t i = 2 * Size - 2; i <= 2 * Size; i -= 2) {
-            out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]);
+            out[i / 2] = static_cast<u8>((ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]));
         }
     } else {
         for (std::size_t i = 0; i < 2 * Size; i += 2) {
-            out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]);
+            out[i / 2] = static_cast<u8>((ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]));
         }
     }
     return out;

src/common/logging/backend.cpp

@@ -274,7 +274,6 @@ const char* GetLogClassName(Class log_class) {
     case Class::Count:
         break;
     }
-    UNREACHABLE();
     return "Invalid";
 }
 
@@ -293,7 +292,6 @@ const char* GetLevelName(Level log_level) {
         break;
     }
 #undef LVL
-    UNREACHABLE();
     return "Invalid";
 }
 

src/common/math_util.h

@@ -20,14 +20,14 @@ struct Rectangle {
 
     constexpr Rectangle() = default;
 
-    constexpr Rectangle(T left, T top, T right, T bottom)
-        : left(left), top(top), right(right), bottom(bottom) {}
+    constexpr Rectangle(T left_, T top_, T right_, T bottom_)
+        : left(left_), top(top_), right(right_), bottom(bottom_) {}
 
     [[nodiscard]] T GetWidth() const {
         if constexpr (std::is_floating_point_v<T>) {
             return std::abs(right - left);
         } else {
-            return std::abs(static_cast<std::make_signed_t<T>>(right - left));
+            return static_cast<T>(std::abs(static_cast<std::make_signed_t<T>>(right - left)));
         }
     }
 
@@ -35,7 +35,7 @@ struct Rectangle {
         if constexpr (std::is_floating_point_v<T>) {
             return std::abs(bottom - top);
         } else {
-            return std::abs(static_cast<std::make_signed_t<T>>(bottom - top));
+            return static_cast<T>(std::abs(static_cast<std::make_signed_t<T>>(bottom - top)));
         }
     }
 

src/common/misc.cpp

@@ -16,16 +16,23 @@
 // Call directly after the command or use the error num.
 // This function might change the error code.
 std::string GetLastErrorMsg() {
-    static const std::size_t buff_size = 255;
+    static constexpr std::size_t buff_size = 255;
     char err_str[buff_size];
 
 #ifdef _WIN32
     FormatMessageA(FORMAT_MESSAGE_FROM_SYSTEM, nullptr, GetLastError(),
                    MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), err_str, buff_size, nullptr);
+    return std::string(err_str, buff_size);
+#elif defined(__GLIBC__) && (_GNU_SOURCE || (_POSIX_C_SOURCE < 200112L && _XOPEN_SOURCE < 600))
+    // Thread safe (GNU-specific)
+    const char* str = strerror_r(errno, err_str, buff_size);
+    return std::string(str);
 #else
     // Thread safe (XSI-compliant)
-    strerror_r(errno, err_str, buff_size);
+    const int success = strerror_r(errno, err_str, buff_size);
+    if (success != 0) {
+        return {};
+    }
+    return std::string(err_str);
 #endif
-
-    return std::string(err_str, buff_size);
 }

src/common/stream.cpp

@@ -0,0 +1,47 @@
+// Copyright 2020 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <stdexcept>
+#include "common/common_types.h"
+#include "common/stream.h"
+
+namespace Common {
+
+Stream::Stream() = default;
+Stream::~Stream() = default;
+
+void Stream::Seek(s32 offset, SeekOrigin origin) {
+    if (origin == SeekOrigin::SetOrigin) {
+        if (offset < 0) {
+            position = 0;
+        } else if (position >= buffer.size()) {
+            position = buffer.size();
+        } else {
+            position = offset;
+        }
+    } else if (origin == SeekOrigin::FromCurrentPos) {
+        Seek(static_cast<s32>(position) + offset, SeekOrigin::SetOrigin);
+    } else if (origin == SeekOrigin::FromEnd) {
+        Seek(static_cast<s32>(buffer.size()) - offset, SeekOrigin::SetOrigin);
+    }
+}
+
+u8 Stream::ReadByte() {
+    if (position < buffer.size()) {
+        return buffer[position++];
+    } else {
+        throw std::out_of_range("Attempting to read a byte not within the buffer range");
+    }
+}
+
+void Stream::WriteByte(u8 byte) {
+    if (position == buffer.size()) {
+        buffer.push_back(byte);
+        position++;
+    } else {
+        buffer.insert(buffer.begin() + position, byte);
+    }
+}
+
+} // namespace Common

src/common/stream.h

@@ -0,0 +1,56 @@
+// Copyright 2020 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <vector>
+#include "common/common_types.h"
+
+namespace Common {
+
+enum class SeekOrigin {
+    SetOrigin,
+    FromCurrentPos,
+    FromEnd,
+};
+
+class Stream {
+public:
+    /// Stream creates a bitstream and provides common functionality on the stream.
+    explicit Stream();
+    ~Stream();
+
+    Stream(const Stream&) = delete;
+    Stream& operator=(const Stream&) = delete;
+
+    Stream(Stream&&) = default;
+    Stream& operator=(Stream&&) = default;
+
+    /// Reposition bitstream "cursor" to the specified offset from origin
+    void Seek(s32 offset, SeekOrigin origin);
+
+    /// Reads next byte in the stream buffer and increments position
+    u8 ReadByte();
+
+    /// Writes byte at current position
+    void WriteByte(u8 byte);
+
+    [[nodiscard]] std::size_t GetPosition() const {
+        return position;
+    }
+
+    [[nodiscard]] std::vector<u8>& GetBuffer() {
+        return buffer;
+    }
+
+    [[nodiscard]] const std::vector<u8>& GetBuffer() const {
+        return buffer;
+    }
+
+private:
+    std::vector<u8> buffer;
+    std::size_t position{0};
+};
+
+} // namespace Common

src/common/string_util.cpp

@@ -8,6 +8,7 @@
 #include <cstdlib>
 #include <locale>
 #include <sstream>
+
 #include "common/common_paths.h"
 #include "common/logging/log.h"
 #include "common/string_util.h"
@@ -21,14 +22,14 @@ namespace Common {
 /// Make a string lowercase
 std::string ToLower(std::string str) {
     std::transform(str.begin(), str.end(), str.begin(),
-                   [](unsigned char c) { return std::tolower(c); });
+                   [](unsigned char c) { return static_cast<char>(std::tolower(c)); });
     return str;
 }
 
 /// Make a string uppercase
 std::string ToUpper(std::string str) {
     std::transform(str.begin(), str.end(), str.begin(),
-                   [](unsigned char c) { return std::toupper(c); });
+                   [](unsigned char c) { return static_cast<char>(std::toupper(c)); });
     return str;
 }
 

src/common/timer.cpp

@@ -142,20 +142,18 @@ std::string Timer::GetTimeFormatted() {
 // ----------------
 double Timer::GetDoubleTime() {
     // Get continuous timestamp
-    u64 TmpSeconds = static_cast<u64>(Common::Timer::GetTimeSinceJan1970().count());
-    double ms = static_cast<u64>(GetTimeMs().count()) % 1000;
+    auto tmp_seconds = static_cast<u64>(GetTimeSinceJan1970().count());
+    const auto ms = static_cast<double>(static_cast<u64>(GetTimeMs().count()) % 1000);
 
     // Remove a few years. We only really want enough seconds to make
     // sure that we are detecting actual actions, perhaps 60 seconds is
     // enough really, but I leave a year of seconds anyway, in case the
     // user's clock is incorrect or something like that.
-    TmpSeconds = TmpSeconds - (38 * 365 * 24 * 60 * 60);
+    tmp_seconds = tmp_seconds - (38 * 365 * 24 * 60 * 60);
 
     // Make a smaller integer that fits in the double
-    u32 Seconds = static_cast<u32>(TmpSeconds);
-    double TmpTime = Seconds + ms;
-
-    return TmpTime;
+    const auto seconds = static_cast<u32>(tmp_seconds);
+    return seconds + ms;
 }
 
 } // Namespace Common

src/common/vector_math.h

@@ -87,7 +87,13 @@ public:
 
     template <typename V>
     [[nodiscard]] constexpr Vec2<decltype(T{} * V{})> operator*(const V& f) const {
-        return {x * f, y * f};
+        using TV = decltype(T{} * V{});
+        using C = std::common_type_t<T, V>;
+
+        return {
+            static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)),
+        };
     }
 
     template <typename V>
@@ -98,7 +104,13 @@ public:
 
     template <typename V>
     [[nodiscard]] constexpr Vec2<decltype(T{} / V{})> operator/(const V& f) const {
-        return {x / f, y / f};
+        using TV = decltype(T{} / V{});
+        using C = std::common_type_t<T, V>;
+
+        return {
+            static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)),
+        };
     }
 
     template <typename V>
@@ -168,7 +180,10 @@ public:
 
 template <typename T, typename V>
 [[nodiscard]] constexpr Vec2<T> operator*(const V& f, const Vec2<T>& vec) {
-    return Vec2<T>(f * vec.x, f * vec.y);
+    using C = std::common_type_t<T, V>;
+
+    return Vec2<T>(static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.x)),
+                   static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.y)));
 }
 
 using Vec2f = Vec2<float>;
@@ -237,7 +252,14 @@ public:
 
     template <typename V>
     [[nodiscard]] constexpr Vec3<decltype(T{} * V{})> operator*(const V& f) const {
-        return {x * f, y * f, z * f};
+        using TV = decltype(T{} * V{});
+        using C = std::common_type_t<T, V>;
+
+        return {
+            static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(z) * static_cast<C>(f)),
+        };
     }
 
     template <typename V>
@@ -247,7 +269,14 @@ public:
     }
     template <typename V>
     [[nodiscard]] constexpr Vec3<decltype(T{} / V{})> operator/(const V& f) const {
-        return {x / f, y / f, z / f};
+        using TV = decltype(T{} / V{});
+        using C = std::common_type_t<T, V>;
+
+        return {
+            static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(z) / static_cast<C>(f)),
+        };
     }
 
     template <typename V>
@@ -367,7 +396,11 @@ public:
 
 template <typename T, typename V>
 [[nodiscard]] constexpr Vec3<T> operator*(const V& f, const Vec3<T>& vec) {
-    return Vec3<T>(f * vec.x, f * vec.y, f * vec.z);
+    using C = std::common_type_t<T, V>;
+
+    return Vec3<T>(static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.x)),
+                   static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.y)),
+                   static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.z)));
 }
 
 template <>
@@ -446,7 +479,15 @@ public:
 
     template <typename V>
     [[nodiscard]] constexpr Vec4<decltype(T{} * V{})> operator*(const V& f) const {
-        return {x * f, y * f, z * f, w * f};
+        using TV = decltype(T{} * V{});
+        using C = std::common_type_t<T, V>;
+
+        return {
+            static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(z) * static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(w) * static_cast<C>(f)),
+        };
     }
 
     template <typename V>
@@ -457,7 +498,15 @@ public:
 
     template <typename V>
     [[nodiscard]] constexpr Vec4<decltype(T{} / V{})> operator/(const V& f) const {
-        return {x / f, y / f, z / f, w / f};
+        using TV = decltype(T{} / V{});
+        using C = std::common_type_t<T, V>;
+
+        return {
+            static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(z) / static_cast<C>(f)),
+            static_cast<TV>(static_cast<C>(w) / static_cast<C>(f)),
+        };
     }
 
     template <typename V>
@@ -582,7 +631,15 @@ public:
 
 template <typename T, typename V>
 [[nodiscard]] constexpr Vec4<decltype(V{} * T{})> operator*(const V& f, const Vec4<T>& vec) {
-    return {f * vec.x, f * vec.y, f * vec.z, f * vec.w};
+    using TV = decltype(V{} * T{});
+    using C = std::common_type_t<T, V>;
+
+    return {
+        static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.x)),
+        static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.y)),
+        static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.z)),
+        static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.w)),
+    };
 }
 
 using Vec4f = Vec4<float>;

src/common/wall_clock.cpp

@@ -53,7 +53,7 @@ public:
         return Common::Divide128On32(temporary, 1000000000).first;
     }
 
-    void Pause(bool is_paused) override {
+    void Pause([[maybe_unused]] bool is_paused) override {
         // Do nothing in this clock type.
     }
 

src/common/x64/native_clock.h

@@ -34,7 +34,7 @@ private:
     /// value used to reduce the native clocks accuracy as some apss rely on
     /// undefined behavior where the level of accuracy in the clock shouldn't
     /// be higher.
-    static constexpr u64 inaccuracy_mask = ~(0x400 - 1);
+    static constexpr u64 inaccuracy_mask = ~(UINT64_C(0x400) - 1);
 
     SpinLock rtsc_serialize{};
     u64 last_measure{};

src/core/CMakeLists.txt

@@ -1,9 +1,3 @@
-if (YUZU_ENABLE_BOXCAT)
-    set(BCAT_BOXCAT_ADDITIONAL_SOURCES hle/service/bcat/backend/boxcat.cpp hle/service/bcat/backend/boxcat.h)
-else()
-    set(BCAT_BOXCAT_ADDITIONAL_SOURCES)
-endif()
-
 add_library(core STATIC
     arm/arm_interface.h
     arm/arm_interface.cpp
@@ -303,7 +297,6 @@ add_library(core STATIC
     hle/service/audio/hwopus.h
     hle/service/bcat/backend/backend.cpp
     hle/service/bcat/backend/backend.h
-    ${BCAT_BOXCAT_ADDITIONAL_SOURCES}
     hle/service/bcat/bcat.cpp
     hle/service/bcat/bcat.h
     hle/service/bcat/module.cpp
@@ -446,6 +439,8 @@ add_library(core STATIC
     hle/service/nvdrv/devices/nvhost_gpu.h
     hle/service/nvdrv/devices/nvhost_nvdec.cpp
     hle/service/nvdrv/devices/nvhost_nvdec.h
+    hle/service/nvdrv/devices/nvhost_nvdec_common.cpp
+    hle/service/nvdrv/devices/nvhost_nvdec_common.h
     hle/service/nvdrv/devices/nvhost_nvjpg.cpp
     hle/service/nvdrv/devices/nvhost_nvjpg.h
     hle/service/nvdrv/devices/nvhost_vic.cpp
@@ -459,6 +454,8 @@ add_library(core STATIC
     hle/service/nvdrv/nvdrv.h
     hle/service/nvdrv/nvmemp.cpp
     hle/service/nvdrv/nvmemp.h
+    hle/service/nvdrv/syncpoint_manager.cpp
+    hle/service/nvdrv/syncpoint_manager.h
     hle/service/nvflinger/buffer_queue.cpp
     hle/service/nvflinger/buffer_queue.h
     hle/service/nvflinger/nvflinger.cpp
@@ -608,6 +605,13 @@ add_library(core STATIC
     tools/freezer.h
 )
 
+if (YUZU_ENABLE_BOXCAT)
+    target_sources(core PRIVATE
+        hle/service/bcat/backend/boxcat.cpp
+        hle/service/bcat/backend/boxcat.h
+    )
+endif()
+
 if (MSVC)
     target_compile_options(core PRIVATE
         # 'expression' : signed/unsigned mismatch
@@ -623,6 +627,20 @@ if (MSVC)
         # 'context' : truncation from 'type1' to 'type2'
         /we4305
     )
+else()
+    target_compile_options(core PRIVATE
+        -Werror=conversion
+        -Werror=ignored-qualifiers
+        -Werror=implicit-fallthrough
+        -Werror=reorder
+        -Werror=sign-compare
+        -Werror=unused-variable
+
+        $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter>
+        $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable>
+
+        -Wno-sign-conversion
+    )
 endif()
 
 create_target_directory_groups(core)

src/core/core.cpp

@@ -40,6 +40,7 @@
 #include "core/hle/service/lm/manager.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/sm/sm.h"
+#include "core/hle/service/time/time_manager.h"
 #include "core/loader/loader.h"
 #include "core/memory.h"
 #include "core/memory/cheat_engine.h"
@@ -121,7 +122,7 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
 struct System::Impl {
     explicit Impl(System& system)
         : kernel{system}, fs_controller{system}, memory{system},
-          cpu_manager{system}, reporter{system}, applet_manager{system} {}
+          cpu_manager{system}, reporter{system}, applet_manager{system}, time_manager{system} {}
 
     ResultStatus Run() {
         status = ResultStatus::Success;
@@ -178,16 +179,21 @@ struct System::Impl {
         arp_manager.ResetAll();
 
         telemetry_session = std::make_unique<Core::TelemetrySession>();
+
+        gpu_core = VideoCore::CreateGPU(emu_window, system);
+        if (!gpu_core) {
+            return ResultStatus::ErrorVideoCore;
+        }
+
         service_manager = std::make_shared<Service::SM::ServiceManager>(kernel);
 
         Service::Init(service_manager, system);
         GDBStub::DeferStart();
 
         interrupt_manager = std::make_unique<Core::Hardware::InterruptManager>(system);
-        gpu_core = VideoCore::CreateGPU(emu_window, system);
-        if (!gpu_core) {
-            return ResultStatus::ErrorVideoCore;
-        }
+
+        // Initialize time manager, which must happen after kernel is created
+        time_manager.Initialize();
 
         is_powered_on = true;
         exit_lock = false;
@@ -387,6 +393,7 @@ struct System::Impl {
     /// Service State
     Service::Glue::ARPManager arp_manager;
     Service::LM::Manager lm_manager{reporter};
+    Service::Time::TimeManager time_manager;
 
     /// Service manager
     std::shared_ptr<Service::SM::ServiceManager> service_manager;
@@ -717,6 +724,14 @@ const Service::LM::Manager& System::GetLogManager() const {
     return impl->lm_manager;
 }
 
+Service::Time::TimeManager& System::GetTimeManager() {
+    return impl->time_manager;
+}
+
+const Service::Time::TimeManager& System::GetTimeManager() const {
+    return impl->time_manager;
+}
+
 void System::SetExitLock(bool locked) {
     impl->exit_lock = locked;
 }

src/core/core.h

@@ -69,6 +69,10 @@ namespace SM {
 class ServiceManager;
 } // namespace SM
 
+namespace Time {
+class TimeManager;
+} // namespace Time
+
 } // namespace Service
 
 namespace Tegra {
@@ -361,6 +365,10 @@ public:
 
     const Service::LM::Manager& GetLogManager() const;
 
+    Service::Time::TimeManager& GetTimeManager();
+
+    const Service::Time::TimeManager& GetTimeManager() const;
+
     void SetExitLock(bool locked);
 
     bool GetExitLock() const;

src/core/cpu_manager.cpp

@@ -365,6 +365,8 @@ void CpuManager::RunThread(std::size_t core) {
     data.enter_barrier.reset();
     data.exit_barrier.reset();
     data.initialized = false;
+
+    MicroProfileOnThreadExit();
 }
 
 } // namespace Core

src/core/crypto/key_manager.cpp

@@ -411,7 +411,7 @@ Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& ke
     // Combine sources and seed
     for (auto& source : sd_key_sources) {
         for (std::size_t i = 0; i < source.size(); ++i) {
-            source[i] ^= sd_seed[i & 0xF];
+            source[i] = static_cast<u8>(source[i] ^ sd_seed[i & 0xF]);
         }
     }
 

src/core/file_sys/fsmitm_romfsbuild.cpp

@@ -266,8 +266,9 @@ std::multimap<u64, VirtualFile> RomFSBuildContext::Build() {
         cur_file->offset = file_partition_size;
         file_partition_size += cur_file->size;
         cur_file->entry_offset = entry_offset;
-        entry_offset += sizeof(RomFSFileEntry) +
-                        Common::AlignUp(cur_file->path_len - cur_file->cur_path_ofs, 4);
+        entry_offset +=
+            static_cast<u32>(sizeof(RomFSFileEntry) +
+                             Common::AlignUp(cur_file->path_len - cur_file->cur_path_ofs, 4));
         prev_file = cur_file;
     }
     // Assign deferred parent/sibling ownership.
@@ -284,8 +285,9 @@ std::multimap<u64, VirtualFile> RomFSBuildContext::Build() {
     for (const auto& it : directories) {
         cur_dir = it.second;
         cur_dir->entry_offset = entry_offset;
-        entry_offset += sizeof(RomFSDirectoryEntry) +
-                        Common::AlignUp(cur_dir->path_len - cur_dir->cur_path_ofs, 4);
+        entry_offset +=
+            static_cast<u32>(sizeof(RomFSDirectoryEntry) +
+                             Common::AlignUp(cur_dir->path_len - cur_dir->cur_path_ofs, 4));
     }
     // Assign deferred parent/sibling ownership.
     for (auto it = directories.rbegin(); it->second != root; ++it) {

src/core/file_sys/ips_layer.cpp

@@ -299,7 +299,7 @@ void IPSwitchCompiler::Parse() {
                              patch_text->GetName(), offset, Common::HexToString(replace));
                 }
 
-                patch.records.insert_or_assign(offset, std::move(replace));
+                patch.records.insert_or_assign(static_cast<u32>(offset), std::move(replace));
             }
 
             patches.push_back(std::move(patch));

src/core/file_sys/nca_metadata.cpp

@@ -108,7 +108,7 @@ std::vector<u8> CNMT::Serialize() const {
         memcpy(out.data() + sizeof(CNMTHeader), &opt_header, sizeof(OptionalHeader));
     }
 
-    auto offset = header.table_offset;
+    u64_le offset = header.table_offset;
 
     for (const auto& rec : content_records) {
         memcpy(out.data() + offset + sizeof(CNMTHeader), &rec, sizeof(ContentRecord));

src/core/file_sys/patch_manager.cpp

@@ -29,7 +29,7 @@
 namespace FileSys {
 namespace {
 
-constexpr u64 SINGLE_BYTE_MODULUS = 0x100;
+constexpr u32 SINGLE_BYTE_MODULUS = 0x100;
 constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000;
 
 constexpr std::array<const char*, 14> EXEFS_FILE_NAMES{

src/core/file_sys/submission_package.cpp

@@ -19,38 +19,6 @@
 #include "core/loader/loader.h"
 
 namespace FileSys {
-namespace {
-void SetTicketKeys(const std::vector<VirtualFile>& files) {
-    auto& keys = Core::Crypto::KeyManager::Instance();
-
-    for (const auto& ticket_file : files) {
-        if (ticket_file == nullptr) {
-            continue;
-        }
-
-        if (ticket_file->GetExtension() != "tik") {
-            continue;
-        }
-
-        if (ticket_file->GetSize() <
-            Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET + sizeof(Core::Crypto::Key128)) {
-            continue;
-        }
-
-        Core::Crypto::Key128 key{};
-        ticket_file->Read(key.data(), key.size(), Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET);
-
-        // We get the name without the extension in order to create the rights ID.
-        std::string name_only(ticket_file->GetName());
-        name_only.erase(name_only.size() - 4);
-
-        const auto rights_id_raw = Common::HexStringToArray<16>(name_only);
-        u128 rights_id;
-        std::memcpy(rights_id.data(), rights_id_raw.data(), sizeof(u128));
-        keys.SetKey(Core::Crypto::S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
-    }
-}
-} // Anonymous namespace
 
 NSP::NSP(VirtualFile file_)
     : file(std::move(file_)), status{Loader::ResultStatus::Success},
@@ -232,6 +200,35 @@ VirtualDir NSP::GetParentDirectory() const {
     return file->GetContainingDirectory();
 }
 
+void NSP::SetTicketKeys(const std::vector<VirtualFile>& files) {
+    for (const auto& ticket_file : files) {
+        if (ticket_file == nullptr) {
+            continue;
+        }
+
+        if (ticket_file->GetExtension() != "tik") {
+            continue;
+        }
+
+        if (ticket_file->GetSize() <
+            Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET + sizeof(Core::Crypto::Key128)) {
+            continue;
+        }
+
+        Core::Crypto::Key128 key{};
+        ticket_file->Read(key.data(), key.size(), Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET);
+
+        // We get the name without the extension in order to create the rights ID.
+        std::string name_only(ticket_file->GetName());
+        name_only.erase(name_only.size() - 4);
+
+        const auto rights_id_raw = Common::HexStringToArray<16>(name_only);
+        u128 rights_id;
+        std::memcpy(rights_id.data(), rights_id_raw.data(), sizeof(u128));
+        keys.SetKey(Core::Crypto::S128KeyType::Titlekey, key, rights_id[1], rights_id[0]);
+    }
+}
+
 void NSP::InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files) {
     exefs = pfs;
 
@@ -286,12 +283,31 @@ void NSP::ReadNCAs(const std::vector<VirtualFile>& files) {
                 }
 
                 auto next_nca = std::make_shared<NCA>(std::move(next_file), nullptr, 0);
+
                 if (next_nca->GetType() == NCAContentType::Program) {
                     program_status[next_nca->GetTitleId()] = next_nca->GetStatus();
                 }
-                if (next_nca->GetStatus() == Loader::ResultStatus::Success ||
-                    (next_nca->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS &&
-                     (next_nca->GetTitleId() & 0x800) != 0)) {
+
+                if (next_nca->GetStatus() != Loader::ResultStatus::Success &&
+                    next_nca->GetStatus() != Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) {
+                    continue;
+                }
+
+                // If the last 3 hexadecimal digits of the CNMT TitleID is 0x800 or is missing the
+                // BKTRBaseRomFS, this is an update NCA. Otherwise, this is a base NCA.
+                if ((cnmt.GetTitleID() & 0x800) != 0 ||
+                    next_nca->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) {
+                    // If the last 3 hexadecimal digits of the NCA's TitleID is between 0x1 and
+                    // 0x7FF, this is a multi-program update NCA. Otherwise, this is a regular
+                    // update NCA.
+                    if ((next_nca->GetTitleId() & 0x7FF) != 0 &&
+                        (next_nca->GetTitleId() & 0x800) == 0) {
+                        ncas[next_nca->GetTitleId()][{cnmt.GetType(), rec.type}] =
+                            std::move(next_nca);
+                    } else {
+                        ncas[cnmt.GetTitleID()][{cnmt.GetType(), rec.type}] = std::move(next_nca);
+                    }
+                } else {
                     ncas[next_nca->GetTitleId()][{cnmt.GetType(), rec.type}] = std::move(next_nca);
                 }
             }

src/core/file_sys/submission_package.h

@@ -63,6 +63,7 @@ public:
     VirtualDir GetParentDirectory() const override;
 
 private:
+    void SetTicketKeys(const std::vector<VirtualFile>& files);
     void InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files);
     void ReadNCAs(const std::vector<VirtualFile>& files);
 

src/core/frontend/applets/controller.cpp

@@ -19,7 +19,7 @@ DefaultControllerApplet::DefaultControllerApplet(Service::SM::ServiceManager& se
 DefaultControllerApplet::~DefaultControllerApplet() = default;
 
 void DefaultControllerApplet::ReconfigureControllers(std::function<void()> callback,
-                                                     ControllerParameters parameters) const {
+                                                     const ControllerParameters& parameters) const {
     LOG_INFO(Service_HID, "called, deducing the best configuration based on the given parameters!");
 
     auto& npad =

src/core/frontend/applets/controller.h

@@ -38,7 +38,7 @@ public:
     virtual ~ControllerApplet();
 
     virtual void ReconfigureControllers(std::function<void()> callback,
-                                        ControllerParameters parameters) const = 0;
+                                        const ControllerParameters& parameters) const = 0;
 };
 
 class DefaultControllerApplet final : public ControllerApplet {
@@ -47,7 +47,7 @@ public:
     ~DefaultControllerApplet() override;
 
     void ReconfigureControllers(std::function<void()> callback,
-                                ControllerParameters parameters) const override;
+                                const ControllerParameters& parameters) const override;
 
 private:
     Service::SM::ServiceManager& service_manager;

src/core/frontend/emu_window.cpp

@@ -84,10 +84,12 @@ void EmuWindow::TouchPressed(unsigned framebuffer_x, unsigned framebuffer_y) {
         return;
 
     std::lock_guard guard{touch_state->mutex};
-    touch_state->touch_x = static_cast<float>(framebuffer_x - framebuffer_layout.screen.left) /
-                           (framebuffer_layout.screen.right - framebuffer_layout.screen.left);
-    touch_state->touch_y = static_cast<float>(framebuffer_y - framebuffer_layout.screen.top) /
-                           (framebuffer_layout.screen.bottom - framebuffer_layout.screen.top);
+    touch_state->touch_x =
+        static_cast<float>(framebuffer_x - framebuffer_layout.screen.left) /
+        static_cast<float>(framebuffer_layout.screen.right - framebuffer_layout.screen.left);
+    touch_state->touch_y =
+        static_cast<float>(framebuffer_y - framebuffer_layout.screen.top) /
+        static_cast<float>(framebuffer_layout.screen.bottom - framebuffer_layout.screen.top);
 
     touch_state->touch_pressed = true;
 }

src/core/frontend/framebuffer_layout.cpp

@@ -14,8 +14,8 @@ namespace Layout {
 template <class T>
 static Common::Rectangle<T> MaxRectangle(Common::Rectangle<T> window_area,
                                          float screen_aspect_ratio) {
-    float scale = std::min(static_cast<float>(window_area.GetWidth()),
-                           window_area.GetHeight() / screen_aspect_ratio);
+    const float scale = std::min(static_cast<float>(window_area.GetWidth()),
+                                 static_cast<float>(window_area.GetHeight()) / screen_aspect_ratio);
     return Common::Rectangle<T>{0, 0, static_cast<T>(std::round(scale)),
                                 static_cast<T>(std::round(scale * screen_aspect_ratio))};
 }
@@ -27,7 +27,7 @@ FramebufferLayout DefaultFrameLayout(u32 width, u32 height) {
     // so just calculate them both even if the other isn't showing.
     FramebufferLayout res{width, height, false, {}};
 
-    const float window_aspect_ratio = static_cast<float>(height) / width;
+    const float window_aspect_ratio = static_cast<float>(height) / static_cast<float>(width);
     const float emulation_aspect_ratio = EmulationAspectRatio(
         static_cast<AspectRatio>(Settings::values.aspect_ratio.GetValue()), window_aspect_ratio);
 

src/core/gdbstub/gdbstub.cpp

@@ -291,11 +291,11 @@ static void FpuWrite(std::size_t id, u128 val, Kernel::Thread* thread = nullptr)
  */
 static u8 HexCharToValue(u8 hex) {
     if (hex >= '0' && hex <= '9') {
-        return hex - '0';
+        return static_cast<u8>(hex - '0');
     } else if (hex >= 'a' && hex <= 'f') {
-        return hex - 'a' + 0xA;
+        return static_cast<u8>(hex - 'a' + 0xA);
     } else if (hex >= 'A' && hex <= 'F') {
-        return hex - 'A' + 0xA;
+        return static_cast<u8>(hex - 'A' + 0xA);
     }
 
     LOG_ERROR(Debug_GDBStub, "Invalid nibble: {} ({:02X})", hex, hex);
@@ -310,9 +310,9 @@ static u8 HexCharToValue(u8 hex) {
 static u8 NibbleToHex(u8 n) {
     n &= 0xF;
     if (n < 0xA) {
-        return '0' + n;
+        return static_cast<u8>('0' + n);
     } else {
-        return 'a' + n - 0xA;
+        return static_cast<u8>('a' + n - 0xA);
     }
 }
 
@@ -355,8 +355,8 @@ static u64 HexToLong(const u8* src, std::size_t len) {
  */
 static void MemToGdbHex(u8* dest, const u8* src, std::size_t len) {
     while (len-- > 0) {
-        u8 tmp = *src++;
-        *dest++ = NibbleToHex(tmp >> 4);
+        const u8 tmp = *src++;
+        *dest++ = NibbleToHex(static_cast<u8>(tmp >> 4));
         *dest++ = NibbleToHex(tmp);
     }
 }
@@ -370,7 +370,7 @@ static void MemToGdbHex(u8* dest, const u8* src, std::size_t len) {
  */
 static void GdbHexToMem(u8* dest, const u8* src, std::size_t len) {
     while (len-- > 0) {
-        *dest++ = (HexCharToValue(src[0]) << 4) | HexCharToValue(src[1]);
+        *dest++ = static_cast<u8>((HexCharToValue(src[0]) << 4) | HexCharToValue(src[1]));
         src += 2;
     }
 }
@@ -602,22 +602,22 @@ static void SendReply(const char* reply) {
 
     memcpy(command_buffer + 1, reply, command_length);
 
-    u8 checksum = CalculateChecksum(command_buffer, command_length + 1);
+    const u8 checksum = CalculateChecksum(command_buffer, command_length + 1);
     command_buffer[0] = GDB_STUB_START;
     command_buffer[command_length + 1] = GDB_STUB_END;
-    command_buffer[command_length + 2] = NibbleToHex(checksum >> 4);
+    command_buffer[command_length + 2] = NibbleToHex(static_cast<u8>(checksum >> 4));
     command_buffer[command_length + 3] = NibbleToHex(checksum);
 
     u8* ptr = command_buffer;
     u32 left = command_length + 4;
     while (left > 0) {
-        int sent_size = send(gdbserver_socket, reinterpret_cast<char*>(ptr), left, 0);
+        const auto sent_size = send(gdbserver_socket, reinterpret_cast<char*>(ptr), left, 0);
         if (sent_size < 0) {
             LOG_ERROR(Debug_GDBStub, "gdb: send failed");
             return Shutdown();
         }
 
-        left -= sent_size;
+        left -= static_cast<u32>(sent_size);
         ptr += sent_size;
     }
 }
@@ -777,10 +777,10 @@ static void ReadCommand() {
         command_buffer[command_length++] = c;
     }
 
-    u8 checksum_received = HexCharToValue(ReadByte()) << 4;
-    checksum_received |= HexCharToValue(ReadByte());
+    auto checksum_received = static_cast<u32>(HexCharToValue(ReadByte()) << 4);
+    checksum_received |= static_cast<u32>(HexCharToValue(ReadByte()));
 
-    u8 checksum_calculated = CalculateChecksum(command_buffer, command_length);
+    const u32 checksum_calculated = CalculateChecksum(command_buffer, command_length);
 
     if (checksum_received != checksum_calculated) {
         LOG_ERROR(Debug_GDBStub,

src/core/hle/ipc_helpers.h

@@ -38,10 +38,11 @@ public:
     explicit RequestHelperBase(Kernel::HLERequestContext& context)
         : context(&context), cmdbuf(context.CommandBuffer()) {}
 
-    void Skip(unsigned size_in_words, bool set_to_null) {
-        if (set_to_null)
+    void Skip(u32 size_in_words, bool set_to_null) {
+        if (set_to_null) {
             memset(cmdbuf + index, 0, size_in_words * sizeof(u32));
-        index += size_in_words;
+        }
+        index += static_cast<ptrdiff_t>(size_in_words);
     }
 
     /**
@@ -49,15 +50,15 @@ public:
      */
     void AlignWithPadding() {
         if (index & 3) {
-            Skip(4 - (index & 3), true);
+            Skip(static_cast<u32>(4 - (index & 3)), true);
         }
     }
 
-    unsigned GetCurrentOffset() const {
-        return static_cast<unsigned>(index);
+    u32 GetCurrentOffset() const {
+        return static_cast<u32>(index);
     }
 
-    void SetCurrentOffset(unsigned offset) {
+    void SetCurrentOffset(u32 offset) {
         index = static_cast<ptrdiff_t>(offset);
     }
 };
@@ -89,7 +90,7 @@ public:
 
         // The entire size of the raw data section in u32 units, including the 16 bytes of mandatory
         // padding.
-        u32 raw_data_size = sizeof(IPC::DataPayloadHeader) / 4 + 4 + normal_params_size;
+        u64 raw_data_size = sizeof(IPC::DataPayloadHeader) / 4 + 4 + normal_params_size;
 
         u32 num_handles_to_move{};
         u32 num_domain_objects{};
@@ -105,7 +106,7 @@ public:
             raw_data_size += sizeof(DomainMessageHeader) / 4 + num_domain_objects;
         }
 
-        header.data_size.Assign(raw_data_size);
+        header.data_size.Assign(static_cast<u32>(raw_data_size));
         if (num_handles_to_copy || num_handles_to_move) {
             header.enable_handle_descriptor.Assign(1);
         }

src/core/hle/kernel/handle_table.cpp

@@ -118,7 +118,7 @@ std::shared_ptr<Object> HandleTable::GetGeneric(Handle handle) const {
 
 void HandleTable::Clear() {
     for (u16 i = 0; i < table_size; ++i) {
-        generations[i] = i + 1;
+        generations[i] = static_cast<u16>(i + 1);
         objects[i] = nullptr;
     }
     next_free_slot = 0;

src/core/hle/kernel/kernel.cpp

@@ -7,7 +7,6 @@
 #include <bitset>
 #include <functional>
 #include <memory>
-#include <mutex>
 #include <thread>
 #include <unordered_map>
 #include <utility>
@@ -87,8 +86,6 @@ struct KernelCore::Impl {
         }
         cores.clear();
 
-        registered_core_threads.reset();
-
         process_list.clear();
         current_process = nullptr;
 
@@ -107,7 +104,11 @@ struct KernelCore::Impl {
         cores.clear();
 
         exclusive_monitor.reset();
-        host_thread_ids.clear();
+
+        num_host_threads = 0;
+        std::fill(register_host_thread_keys.begin(), register_host_thread_keys.end(),
+                  std::thread::id{});
+        std::fill(register_host_thread_values.begin(), register_host_thread_values.end(), 0);
     }
 
     void InitializePhysicalCores() {
@@ -177,54 +178,58 @@ struct KernelCore::Impl {
 
     void MakeCurrentProcess(Process* process) {
         current_process = process;
-
         if (process == nullptr) {
             return;
         }
-
-        u32 core_id = GetCurrentHostThreadID();
+        const u32 core_id = GetCurrentHostThreadID();
         if (core_id < Core::Hardware::NUM_CPU_CORES) {
             system.Memory().SetCurrentPageTable(*process, core_id);
         }
     }
 
     void RegisterCoreThread(std::size_t core_id) {
-        std::unique_lock lock{register_thread_mutex};
-        if (!is_multicore) {
-            single_core_thread_id = std::this_thread::get_id();
-        }
         const std::thread::id this_id = std::this_thread::get_id();
-        const auto it = host_thread_ids.find(this_id);
+        if (!is_multicore) {
+            single_core_thread_id = this_id;
+        }
+        const auto end =
+            register_host_thread_keys.begin() + static_cast<ptrdiff_t>(num_host_threads);
+        const auto it = std::find(register_host_thread_keys.begin(), end, this_id);
         ASSERT(core_id < Core::Hardware::NUM_CPU_CORES);
-        ASSERT(it == host_thread_ids.end());
-        ASSERT(!registered_core_threads[core_id]);
-        host_thread_ids[this_id] = static_cast<u32>(core_id);
-        registered_core_threads.set(core_id);
+        ASSERT(it == end);
+        InsertHostThread(static_cast<u32>(core_id));
     }
 
     void RegisterHostThread() {
-        std::unique_lock lock{register_thread_mutex};
         const std::thread::id this_id = std::this_thread::get_id();
-        const auto it = host_thread_ids.find(this_id);
-        if (it != host_thread_ids.end()) {
-            return;
+        const auto end =
+            register_host_thread_keys.begin() + static_cast<ptrdiff_t>(num_host_threads);
+        const auto it = std::find(register_host_thread_keys.begin(), end, this_id);
+        if (it == end) {
+            InsertHostThread(registered_thread_ids++);
         }
-        host_thread_ids[this_id] = registered_thread_ids++;
     }
 
-    u32 GetCurrentHostThreadID() const {
+    void InsertHostThread(u32 value) {
+        const size_t index = num_host_threads++;
+        ASSERT_MSG(index < NUM_REGISTRABLE_HOST_THREADS, "Too many host threads");
+        register_host_thread_values[index] = value;
+        register_host_thread_keys[index] = std::this_thread::get_id();
+    }
+
+    [[nodiscard]] u32 GetCurrentHostThreadID() const {
         const std::thread::id this_id = std::this_thread::get_id();
-        if (!is_multicore) {
-            if (single_core_thread_id == this_id) {
-                return static_cast<u32>(system.GetCpuManager().CurrentCore());
-            }
+        if (!is_multicore && single_core_thread_id == this_id) {
+            return static_cast<u32>(system.GetCpuManager().CurrentCore());
         }
-        std::unique_lock lock{register_thread_mutex};
-        const auto it = host_thread_ids.find(this_id);
-        if (it == host_thread_ids.end()) {
+        const auto end =
+            register_host_thread_keys.begin() + static_cast<ptrdiff_t>(num_host_threads);
+        const auto it = std::find(register_host_thread_keys.begin(), end, this_id);
+        if (it == end) {
             return Core::INVALID_HOST_THREAD_ID;
         }
-        return it->second;
+        return register_host_thread_values[static_cast<size_t>(
+            std::distance(register_host_thread_keys.begin(), it))];
     }
 
     Core::EmuThreadHandle GetCurrentEmuThreadID() const {
@@ -322,10 +327,14 @@ struct KernelCore::Impl {
     std::vector<Kernel::PhysicalCore> cores;
 
     // 0-3 IDs represent core threads, >3 represent others
-    std::unordered_map<std::thread::id, u32> host_thread_ids;
-    u32 registered_thread_ids{Core::Hardware::NUM_CPU_CORES};
-    std::bitset<Core::Hardware::NUM_CPU_CORES> registered_core_threads;
-    mutable std::mutex register_thread_mutex;
+    std::atomic<u32> registered_thread_ids{Core::Hardware::NUM_CPU_CORES};
+
+    // Number of host threads is a relatively high number to avoid overflowing
+    static constexpr size_t NUM_REGISTRABLE_HOST_THREADS = 64;
+    std::atomic<size_t> num_host_threads{0};
+    std::array<std::atomic<std::thread::id>, NUM_REGISTRABLE_HOST_THREADS>
+        register_host_thread_keys{};
+    std::array<std::atomic<u32>, NUM_REGISTRABLE_HOST_THREADS> register_host_thread_values{};
 
     // Kernel memory management
     std::unique_ptr<Memory::MemoryManager> memory_manager;

src/core/hle/kernel/process.cpp

@@ -4,6 +4,7 @@
 
 #include <algorithm>
 #include <bitset>
+#include <ctime>
 #include <memory>
 #include <random>
 #include "common/alignment.h"
@@ -123,7 +124,7 @@ std::shared_ptr<Process> Process::Create(Core::System& system, std::string name,
                                                               : kernel.CreateNewUserProcessID();
     process->capabilities.InitializeForMetadatalessProcess();
 
-    std::mt19937 rng(Settings::values.rng_seed.GetValue().value_or(0));
+    std::mt19937 rng(Settings::values.rng_seed.GetValue().value_or(std::time(nullptr)));
     std::uniform_int_distribution<u64> distribution;
     std::generate(process->random_entropy.begin(), process->random_entropy.end(),
                   [&] { return distribution(rng); });

src/core/hle/kernel/scheduler.cpp

@@ -72,7 +72,7 @@ u32 GlobalScheduler::SelectThreads() {
         if (top_thread != nullptr) {
             // TODO(Blinkhawk): Implement Thread Pinning
         } else {
-            idle_cores |= (1ul << core);
+            idle_cores |= (1U << core);
         }
         top_threads[core] = top_thread;
     }
@@ -126,7 +126,7 @@ u32 GlobalScheduler::SelectThreads() {
             top_threads[core_id] = suggested;
         }
 
-        idle_cores &= ~(1ul << core_id);
+        idle_cores &= ~(1U << core_id);
     }
     u32 cores_needing_context_switch{};
     for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) {
@@ -134,7 +134,7 @@ u32 GlobalScheduler::SelectThreads() {
         ASSERT(top_threads[core] == nullptr ||
                static_cast<u32>(top_threads[core]->GetProcessorID()) == core);
         if (update_thread(top_threads[core], sched)) {
-            cores_needing_context_switch |= (1ul << core);
+            cores_needing_context_switch |= (1U << core);
         }
     }
     return cores_needing_context_switch;
@@ -364,7 +364,7 @@ void GlobalScheduler::EnableInterruptAndSchedule(u32 cores_pending_reschedule,
         } else {
             must_context_switch = true;
         }
-        cores_pending_reschedule &= ~(1ul << core);
+        cores_pending_reschedule &= ~(1U << core);
     }
     if (must_context_switch) {
         auto& core_scheduler = kernel.CurrentScheduler();
@@ -767,7 +767,7 @@ void Scheduler::SwitchToCurrent() {
                     current_thread->context_guard.unlock();
                     break;
                 }
-                if (current_thread->GetProcessorID() != core_id) {
+                if (static_cast<u32>(current_thread->GetProcessorID()) != core_id) {
                     current_thread->context_guard.unlock();
                     break;
                 }

src/core/hle/service/acc/acc.cpp

@@ -496,7 +496,7 @@ public:
             {3, nullptr, "LoadIdTokenCache"},
             {130, nullptr, "GetNintendoAccountUserResourceCacheForApplication"},
             {150, nullptr, "CreateAuthorizationRequest"},
-            {160, nullptr, "StoreOpenContext"},
+            {160, &IManagerForApplication::StoreOpenContext, "StoreOpenContext"},
             {170, nullptr, "LoadNetworkServiceLicenseKindAsync"},
         };
         // clang-format on
@@ -520,6 +520,12 @@ private:
         rb.PushRaw<u64>(user_id.GetNintendoID());
     }
 
+    void StoreOpenContext(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_ACC, "(STUBBED) called");
+        IPC::ResponseBuilder rb{ctx, 2};
+        rb.Push(RESULT_SUCCESS);
+    }
+
     Common::UUID user_id;
 };
 

src/core/hle/service/am/am.cpp

@@ -1201,6 +1201,8 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_)
         {151, nullptr, "TryPopFromNotificationStorageChannel"},
         {160, nullptr, "GetHealthWarningDisappearedSystemEvent"},
         {170, nullptr, "SetHdcpAuthenticationActivated"},
+        {180, nullptr, "GetLaunchRequiredVersion"},
+        {181, nullptr, "UpgradeLaunchRequiredVersion"},
         {500, nullptr, "StartContinuousRecordingFlushForDebug"},
         {1000, nullptr, "CreateMovieMaker"},
         {1001, nullptr, "PrepareForJit"},

src/core/hle/service/bcat/backend/backend.cpp

@@ -84,7 +84,7 @@ void ProgressServiceBackend::FinishDownload(ResultCode result) {
 
 void ProgressServiceBackend::SignalUpdate() const {
     if (need_hle_lock) {
-        std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
+        std::lock_guard lock(HLE::g_hle_lock);
         event.writable->Signal();
     } else {
         event.writable->Signal();

src/core/hle/service/bcat/backend/boxcat.cpp

@@ -196,7 +196,9 @@ private:
                                     const std::string& content_type_name) {
         if (client == nullptr) {
             client = std::make_unique<httplib::SSLClient>(BOXCAT_HOSTNAME, PORT);
-            client->set_timeout_sec(timeout_seconds);
+            client->set_connection_timeout(timeout_seconds);
+            client->set_read_timeout(timeout_seconds);
+            client->set_write_timeout(timeout_seconds);
         }
 
         httplib::Headers headers{
@@ -255,7 +257,7 @@ private:
         return out;
     }
 
-    std::unique_ptr<httplib::Client> client;
+    std::unique_ptr<httplib::SSLClient> client;
     std::string path;
     u64 title_id;
     u64 build_id;
@@ -443,13 +445,25 @@ std::optional<std::vector<u8>> Boxcat::GetLaunchParameter(TitleIDVersion title)
 Boxcat::StatusResult Boxcat::GetStatus(std::optional<std::string>& global,
                                        std::map<std::string, EventStatus>& games) {
     httplib::SSLClient client{BOXCAT_HOSTNAME, static_cast<int>(PORT)};
-    client.set_timeout_sec(static_cast<int>(TIMEOUT_SECONDS));
+    client.set_connection_timeout(static_cast<int>(TIMEOUT_SECONDS));
+    client.set_read_timeout(static_cast<int>(TIMEOUT_SECONDS));
+    client.set_write_timeout(static_cast<int>(TIMEOUT_SECONDS));
 
     httplib::Headers headers{
         {std::string("Game-Assets-API-Version"), std::string(BOXCAT_API_VERSION)},
         {std::string("Boxcat-Client-Type"), std::string(BOXCAT_CLIENT_TYPE)},
     };
 
+    if (!client.is_valid()) {
+        LOG_ERROR(Service_BCAT, "Client is invalid, going offline!");
+        return StatusResult::Offline;
+    }
+
+    if (!client.is_socket_open()) {
+        LOG_ERROR(Service_BCAT, "Failed to open socket, going offline!");
+        return StatusResult::Offline;
+    }
+
     const auto response = client.Get(BOXCAT_PATHNAME_EVENTS, headers);
     if (response == nullptr)
         return StatusResult::Offline;

src/core/hle/service/caps/caps_c.cpp

@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include "common/logging/log.h"
+#include "core/hle/ipc_helpers.h"
 #include "core/hle/service/caps/caps_c.h"
 
 namespace Service::Capture {
@@ -47,7 +49,7 @@ CAPS_C::CAPS_C() : ServiceFramework("caps:c") {
     static const FunctionInfo functions[] = {
         {1, nullptr, "CaptureRawImage"},
         {2, nullptr, "CaptureRawImageWithTimeout"},
-        {33, nullptr, "Unknown33"},
+        {33, &CAPS_C::SetShimLibraryVersion, "SetShimLibraryVersion"},
         {1001, nullptr, "RequestTakingScreenShot"},
         {1002, nullptr, "RequestTakingScreenShotWithTimeout"},
         {1011, nullptr, "NotifyTakingScreenShotRefused"},
@@ -72,4 +74,16 @@ CAPS_C::CAPS_C() : ServiceFramework("caps:c") {
 
 CAPS_C::~CAPS_C() = default;
 
+void CAPS_C::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto library_version{rp.Pop<u64>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}",
+                library_version, applet_resource_user_id);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 } // namespace Service::Capture

src/core/hle/service/caps/caps_c.h

@@ -16,6 +16,9 @@ class CAPS_C final : public ServiceFramework<CAPS_C> {
 public:
     explicit CAPS_C();
     ~CAPS_C() override;
+
+private:
+    void SetShimLibraryVersion(Kernel::HLERequestContext& ctx);
 };
 
 } // namespace Service::Capture

src/core/hle/service/caps/caps_su.cpp

@@ -25,7 +25,12 @@ CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") {
 CAPS_SU::~CAPS_SU() = default;
 
 void CAPS_SU::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) {
-    LOG_WARNING(Service_Capture, "(STUBBED) called");
+    IPC::RequestParser rp{ctx};
+    const auto library_version{rp.Pop<u64>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}",
+                library_version, applet_resource_user_id);
 
     IPC::ResponseBuilder rb{ctx, 2};
     rb.Push(RESULT_SUCCESS);

src/core/hle/service/caps/caps_u.cpp

@@ -31,8 +31,7 @@ public:
 CAPS_U::CAPS_U() : ServiceFramework("caps:u") {
     // clang-format off
     static const FunctionInfo functions[] = {
-        {31, nullptr, "GetShimLibraryVersion"},
-        {32, nullptr, "SetShimLibraryVersion"},
+        {32, &CAPS_U::SetShimLibraryVersion, "SetShimLibraryVersion"},
         {102, &CAPS_U::GetAlbumContentsFileListForApplication, "GetAlbumContentsFileListForApplication"},
         {103, nullptr, "DeleteAlbumContentsFileForApplication"},
         {104, nullptr, "GetAlbumContentsFileSizeForApplication"},
@@ -53,6 +52,18 @@ CAPS_U::CAPS_U() : ServiceFramework("caps:u") {
 
 CAPS_U::~CAPS_U() = default;
 
+void CAPS_U::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto library_version{rp.Pop<u64>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}",
+                library_version, applet_resource_user_id);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 void CAPS_U::GetAlbumContentsFileListForApplication(Kernel::HLERequestContext& ctx) {
     // Takes a type-0x6 output buffer containing an array of ApplicationAlbumFileEntry, a PID, an
     // u8 ContentType, two s64s, and an u64 AppletResourceUserId. Returns an output u64 for total

src/core/hle/service/caps/caps_u.h

@@ -18,6 +18,7 @@ public:
     ~CAPS_U() override;
 
 private:
+    void SetShimLibraryVersion(Kernel::HLERequestContext& ctx);
     void GetAlbumContentsFileListForApplication(Kernel::HLERequestContext& ctx);
 };
 

src/core/hle/service/filesystem/filesystem.cpp

@@ -79,7 +79,7 @@ ResultCode VfsDirectoryServiceWrapper::DeleteFile(const std::string& path_) cons
     }
 
     auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path));
-    if (dir->GetFile(Common::FS::GetFilename(path)) == nullptr) {
+    if (dir == nullptr || dir->GetFile(Common::FS::GetFilename(path)) == nullptr) {
         return FileSys::ERROR_PATH_NOT_FOUND;
     }
     if (!dir->DeleteFile(Common::FS::GetFilename(path))) {
@@ -93,8 +93,9 @@ ResultCode VfsDirectoryServiceWrapper::DeleteFile(const std::string& path_) cons
 ResultCode VfsDirectoryServiceWrapper::CreateDirectory(const std::string& path_) const {
     std::string path(Common::FS::SanitizePath(path_));
     auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path));
-    if (dir == nullptr && Common::FS::GetFilename(Common::FS::GetParentPath(path)).empty())
+    if (dir == nullptr || Common::FS::GetFilename(Common::FS::GetParentPath(path)).empty()) {
         dir = backing;
+    }
     auto new_dir = dir->CreateSubdirectory(Common::FS::GetFilename(path));
     if (new_dir == nullptr) {
         // TODO(DarkLordZach): Find a better error code for this

src/core/hle/service/hid/controllers/keyboard.cpp

@@ -42,8 +42,8 @@ void Controller_Keyboard::OnUpdate(const Core::Timing::CoreTiming& core_timing,
     cur_entry.modifier = 0;
     if (Settings::values.keyboard_enabled) {
         for (std::size_t i = 0; i < keyboard_keys.size(); ++i) {
-            cur_entry.key[i / KEYS_PER_BYTE] |=
-                (keyboard_keys[i]->GetStatus() << (i % KEYS_PER_BYTE));
+            auto& entry = cur_entry.key[i / KEYS_PER_BYTE];
+            entry = static_cast<u8>(entry | (keyboard_keys[i]->GetStatus() << (i % KEYS_PER_BYTE)));
         }
 
         for (std::size_t i = 0; i < keyboard_mods.size(); ++i) {

src/core/hle/service/hid/controllers/npad.cpp

@@ -269,7 +269,6 @@ void Controller_NPad::RequestPadStateUpdate(u32 npad_id) {
     auto& rstick_entry = npad_pad_states[controller_idx].r_stick;
     const auto& button_state = buttons[controller_idx];
     const auto& analog_state = sticks[controller_idx];
-    const auto& motion_state = motions[controller_idx];
     const auto [stick_l_x_f, stick_l_y_f] =
         analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetStatus();
     const auto [stick_r_x_f, stick_r_y_f] =
@@ -391,18 +390,6 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8*
 
         libnx_entry.connection_status.raw = 0;
         libnx_entry.connection_status.IsConnected.Assign(1);
-        auto& full_sixaxis_entry =
-            npad.sixaxis_full.sixaxis[npad.sixaxis_full.common.last_entry_index];
-        auto& handheld_sixaxis_entry =
-            npad.sixaxis_handheld.sixaxis[npad.sixaxis_handheld.common.last_entry_index];
-        auto& dual_left_sixaxis_entry =
-            npad.sixaxis_dual_left.sixaxis[npad.sixaxis_dual_left.common.last_entry_index];
-        auto& dual_right_sixaxis_entry =
-            npad.sixaxis_dual_right.sixaxis[npad.sixaxis_dual_right.common.last_entry_index];
-        auto& left_sixaxis_entry =
-            npad.sixaxis_left.sixaxis[npad.sixaxis_left.common.last_entry_index];
-        auto& right_sixaxis_entry =
-            npad.sixaxis_right.sixaxis[npad.sixaxis_right.common.last_entry_index];
 
         switch (controller_type) {
         case NPadControllerType::None:
@@ -541,18 +528,6 @@ void Controller_NPad::OnMotionUpdate(const Core::Timing::CoreTiming& core_timing
             }
         }
 
-        auto& main_controller =
-            npad.main_controller_states.npad[npad.main_controller_states.common.last_entry_index];
-        auto& handheld_entry =
-            npad.handheld_states.npad[npad.handheld_states.common.last_entry_index];
-        auto& dual_entry = npad.dual_states.npad[npad.dual_states.common.last_entry_index];
-        auto& left_entry = npad.left_joy_states.npad[npad.left_joy_states.common.last_entry_index];
-        auto& right_entry =
-            npad.right_joy_states.npad[npad.right_joy_states.common.last_entry_index];
-        auto& pokeball_entry =
-            npad.pokeball_states.npad[npad.pokeball_states.common.last_entry_index];
-        auto& libnx_entry = npad.libnx.npad[npad.libnx.common.last_entry_index];
-
         auto& full_sixaxis_entry =
             npad.sixaxis_full.sixaxis[npad.sixaxis_full.common.last_entry_index];
         auto& handheld_sixaxis_entry =
@@ -846,6 +821,15 @@ Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) {
     }
 }
 
+bool Controller_NPad::IsUnintendedHomeButtonInputProtectionEnabled(u32 npad_id) const {
+    return unintended_home_button_input_protection[NPadIdToIndex(npad_id)];
+}
+
+void Controller_NPad::SetUnintendedHomeButtonInputProtectionEnabled(bool is_protection_enabled,
+                                                                    u32 npad_id) {
+    unintended_home_button_input_protection[NPadIdToIndex(npad_id)] = is_protection_enabled;
+}
+
 void Controller_NPad::SetVibrationEnabled(bool can_vibrate) {
     can_controllers_vibrate = can_vibrate;
 }

src/core/hle/service/hid/controllers/npad.h

@@ -146,6 +146,8 @@ public:
     bool IsSixAxisSensorAtRest() const;
     void SetSixAxisEnabled(bool six_axis_status);
     LedPattern GetLedPattern(u32 npad_id);
+    bool IsUnintendedHomeButtonInputProtectionEnabled(u32 npad_id) const;
+    void SetUnintendedHomeButtonInputProtectionEnabled(bool is_protection_enabled, u32 npad_id);
     void SetVibrationEnabled(bool can_vibrate);
     bool IsVibrationEnabled() const;
     void ClearAllConnectedControllers();
@@ -387,6 +389,7 @@ private:
     std::array<Kernel::EventPair, 10> styleset_changed_events;
     Vibration last_processed_vibration{};
     std::array<ControllerHolder, 10> connected_controllers{};
+    std::array<bool, 10> unintended_home_button_input_protection{};
     GyroscopeZeroDriftMode gyroscope_zero_drift_mode{GyroscopeZeroDriftMode::Standard};
     bool can_controllers_vibrate{true};
     bool sixaxis_sensors_enabled{true};

src/core/hle/service/hid/hid.cpp

@@ -224,8 +224,8 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) {
         {128, &Hid::SetNpadHandheldActivationMode, "SetNpadHandheldActivationMode"},
         {129, &Hid::GetNpadHandheldActivationMode, "GetNpadHandheldActivationMode"},
         {130, &Hid::SwapNpadAssignment, "SwapNpadAssignment"},
-        {131, nullptr, "IsUnintendedHomeButtonInputProtectionEnabled"},
-        {132, nullptr, "EnableUnintendedHomeButtonInputProtection"},
+        {131, &Hid::IsUnintendedHomeButtonInputProtectionEnabled, "IsUnintendedHomeButtonInputProtectionEnabled"},
+        {132, &Hid::EnableUnintendedHomeButtonInputProtection, "EnableUnintendedHomeButtonInputProtection"},
         {133, nullptr, "SetNpadJoyAssignmentModeSingleWithDestination"},
         {134, nullptr, "SetNpadAnalogStickUseCenterClamp"},
         {135, nullptr, "SetNpadCaptureButtonAssignment"},
@@ -260,7 +260,7 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) {
         {404, nullptr, "HasLeftRightBattery"},
         {405, nullptr, "GetNpadInterfaceType"},
         {406, nullptr, "GetNpadLeftRightInterfaceType"},
-        {407, nullptr, "GetNpadOfHighestBatteryLevelForJoyLeft"},
+        {407, nullptr, "GetNpadOfHighestBatteryLevel"},
         {408, nullptr, "GetNpadOfHighestBatteryLevelForJoyRight"},
         {500, nullptr, "GetPalmaConnectionHandle"},
         {501, nullptr, "InitializePalma"},
@@ -475,7 +475,7 @@ void Hid::StopSixAxisSensor(Kernel::HLERequestContext& ctx) {
 
 void Hid::EnableSixAxisSensorFusion(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
-    const auto enable{rp.Pop<bool>()};
+    [[maybe_unused]] const auto enable{rp.Pop<bool>()};
     const auto handle{rp.Pop<u32>()};
     const auto applet_resource_user_id{rp.Pop<u64>()};
 
@@ -809,6 +809,40 @@ void Hid::SwapNpadAssignment(Kernel::HLERequestContext& ctx) {
     }
 }
 
+void Hid::IsUnintendedHomeButtonInputProtectionEnabled(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto npad_id{rp.Pop<u32>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", npad_id,
+                applet_resource_user_id);
+
+    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+
+    IPC::ResponseBuilder rb{ctx, 3};
+    rb.Push(RESULT_SUCCESS);
+    rb.Push<bool>(controller.IsUnintendedHomeButtonInputProtectionEnabled(npad_id));
+}
+
+void Hid::EnableUnintendedHomeButtonInputProtection(Kernel::HLERequestContext& ctx) {
+    IPC::RequestParser rp{ctx};
+    const auto unintended_home_button_input_protection{rp.Pop<bool>()};
+    const auto npad_id{rp.Pop<u32>()};
+    const auto applet_resource_user_id{rp.Pop<u64>()};
+
+    LOG_WARNING(Service_HID,
+                "(STUBBED) called, unintended_home_button_input_protection={}, npad_id={},"
+                "applet_resource_user_id={}",
+                npad_id, unintended_home_button_input_protection, applet_resource_user_id);
+
+    auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad);
+    controller.SetUnintendedHomeButtonInputProtectionEnabled(
+        unintended_home_button_input_protection, npad_id);
+
+    IPC::ResponseBuilder rb{ctx, 2};
+    rb.Push(RESULT_SUCCESS);
+}
+
 void Hid::BeginPermitVibrationSession(Kernel::HLERequestContext& ctx) {
     IPC::RequestParser rp{ctx};
     const auto applet_resource_user_id{rp.Pop<u64>()};

src/core/hle/service/hid/hid.h

@@ -123,6 +123,8 @@ private:
     void SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx);
     void GetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx);
     void SwapNpadAssignment(Kernel::HLERequestContext& ctx);
+    void IsUnintendedHomeButtonInputProtectionEnabled(Kernel::HLERequestContext& ctx);
+    void EnableUnintendedHomeButtonInputProtection(Kernel::HLERequestContext& ctx);
     void BeginPermitVibrationSession(Kernel::HLERequestContext& ctx);
     void EndPermitVibrationSession(Kernel::HLERequestContext& ctx);
     void SendVibrationValue(Kernel::HLERequestContext& ctx);

src/core/hle/service/ldr/ldr.cpp

@@ -23,7 +23,7 @@ namespace Service::LDR {
 
 constexpr ResultCode ERROR_INSUFFICIENT_ADDRESS_SPACE{ErrorModule::RO, 2};
 
-constexpr ResultCode ERROR_INVALID_MEMORY_STATE{ErrorModule::Loader, 51};
+[[maybe_unused]] constexpr ResultCode ERROR_INVALID_MEMORY_STATE{ErrorModule::Loader, 51};
 constexpr ResultCode ERROR_INVALID_NRO{ErrorModule::Loader, 52};
 constexpr ResultCode ERROR_INVALID_NRR{ErrorModule::Loader, 53};
 constexpr ResultCode ERROR_MISSING_NRR_HASH{ErrorModule::Loader, 54};
@@ -33,7 +33,7 @@ constexpr ResultCode ERROR_ALREADY_LOADED{ErrorModule::Loader, 57};
 constexpr ResultCode ERROR_INVALID_ALIGNMENT{ErrorModule::Loader, 81};
 constexpr ResultCode ERROR_INVALID_SIZE{ErrorModule::Loader, 82};
 constexpr ResultCode ERROR_INVALID_NRO_ADDRESS{ErrorModule::Loader, 84};
-constexpr ResultCode ERROR_INVALID_NRR_ADDRESS{ErrorModule::Loader, 85};
+[[maybe_unused]] constexpr ResultCode ERROR_INVALID_NRR_ADDRESS{ErrorModule::Loader, 85};
 constexpr ResultCode ERROR_NOT_INITIALIZED{ErrorModule::Loader, 87};
 
 constexpr std::size_t MAXIMUM_LOADED_RO{0x40};
@@ -166,7 +166,7 @@ public:
             {0, &RelocatableObject::LoadNro, "LoadNro"},
             {1, &RelocatableObject::UnloadNro, "UnloadNro"},
             {2, &RelocatableObject::LoadNrr, "LoadNrr"},
-            {3, nullptr, "UnloadNrr"},
+            {3, &RelocatableObject::UnloadNrr, "UnloadNrr"},
             {4, &RelocatableObject::Initialize, "Initialize"},
             {10, nullptr, "LoadNrrEx"},
         };
@@ -272,6 +272,20 @@ public:
         rb.Push(RESULT_SUCCESS);
     }
 
+    void UnloadNrr(Kernel::HLERequestContext& ctx) {
+        IPC::RequestParser rp{ctx};
+        const auto pid = rp.Pop<u64>();
+        const auto nrr_address = rp.Pop<VAddr>();
+
+        LOG_DEBUG(Service_LDR, "called with pid={}, nrr_address={:016X}", pid, nrr_address);
+
+        nrr.erase(nrr_address);
+
+        IPC::ResponseBuilder rb{ctx, 2};
+
+        rb.Push(RESULT_SUCCESS);
+    }
+
     bool ValidateRegionForMap(Kernel::Memory::PageTable& page_table, VAddr start,
                               std::size_t size) const {
         constexpr std::size_t padding_size{4 * Kernel::Memory::PageSize};

src/core/hle/service/mii/manager.cpp

@@ -131,7 +131,7 @@ template <typename T>
 T GetRandomValue(T min, T max) {
     std::random_device device;
     std::mt19937 gen(device());
-    std::uniform_int_distribution<u64> distribution(0, static_cast<u64>(max));
+    std::uniform_int_distribution<u64> distribution(static_cast<u64>(min), static_cast<u64>(max));
     return static_cast<T>(distribution(gen));
 }
 
@@ -428,7 +428,7 @@ bool MiiManager::IsFullDatabase() const {
 }
 
 u32 MiiManager::GetCount(SourceFlag source_flag) const {
-    u32 count{};
+    std::size_t count{};
     if ((source_flag & SourceFlag::Database) != SourceFlag::None) {
         // TODO(bunnei): We don't implement the Mii database, but when we do, update this
         count += 0;
@@ -436,7 +436,7 @@ u32 MiiManager::GetCount(SourceFlag source_flag) const {
     if ((source_flag & SourceFlag::Default) != SourceFlag::None) {
         count += DefaultMiiCount;
     }
-    return count;
+    return static_cast<u32>(count);
 }
 
 ResultVal<MiiInfo> MiiManager::UpdateLatest([[maybe_unused]] const MiiInfo& info,

src/core/hle/service/mii/mii.cpp

@@ -47,6 +47,7 @@ public:
             {23, nullptr, "Convert"},
             {24, nullptr, "ConvertCoreDataToCharInfo"},
             {25, nullptr, "ConvertCharInfoToCoreData"},
+            {26, nullptr, "Append"},
         };
         // clang-format on
 

src/core/hle/service/nifm/nifm.cpp

@@ -62,7 +62,7 @@ public:
             {18, nullptr, "SetRequirementByRevision"},
             {19, nullptr, "GetRequirement"},
             {20, nullptr, "GetRevision"},
-            {21, nullptr, "GetAppletInfo"},
+            {21, &IRequest::GetAppletInfo, "GetAppletInfo"},
             {22, nullptr, "GetAdditionalInfo"},
             {23, nullptr, "SetKeptInSleep"},
             {24, nullptr, "RegisterSocketDescriptor"},
@@ -125,6 +125,16 @@ private:
         rb.Push(RESULT_SUCCESS);
     }
 
+    void GetAppletInfo(Kernel::HLERequestContext& ctx) {
+        LOG_WARNING(Service_NIFM, "(STUBBED) called");
+
+        IPC::ResponseBuilder rb{ctx, 8};
+        rb.Push(RESULT_SUCCESS);
+        rb.Push<u32>(0);
+        rb.Push<u32>(0);
+        rb.Push<u32>(0);
+    }
+
     Kernel::EventPair event1, event2;
 };
 

src/core/hle/service/ns/pl_u.cpp

@@ -50,19 +50,9 @@ constexpr std::array<std::pair<FontArchives, const char*>, 7> SHARED_FONTS{
     std::make_pair(FontArchives::Extension, "nintendo_ext2_003.bfttf"),
 };
 
-constexpr std::array<const char*, 7> SHARED_FONTS_TTF{
-    "FontStandard.ttf",
-    "FontChineseSimplified.ttf",
-    "FontExtendedChineseSimplified.ttf",
-    "FontChineseTraditional.ttf",
-    "FontKorean.ttf",
-    "FontNintendoExtended.ttf",
-    "FontNintendoExtended2.ttf",
-};
-
 // The below data is specific to shared font data dumped from Switch on f/w 2.2
 // Virtual address and offsets/sizes likely will vary by dump
-constexpr VAddr SHARED_FONT_MEM_VADDR{0x00000009d3016000ULL};
+[[maybe_unused]] constexpr VAddr SHARED_FONT_MEM_VADDR{0x00000009d3016000ULL};
 constexpr u32 EXPECTED_RESULT{0x7f9a0218}; // What we expect the decrypted bfttf first 4 bytes to be
 constexpr u32 EXPECTED_MAGIC{0x36f81a1e};  // What we expect the encrypted bfttf first 4 bytes to be
 constexpr u64 SHARED_FONT_MEM_SIZE{0x1100000};

src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp

@@ -46,6 +46,8 @@ u32 nvhost_as_gpu::ioctl(Ioctl command, const std::vector<u8>& input, const std:
         return GetVARegions(input, output);
     case IoctlCommand::IocUnmapBufferCommand:
         return UnmapBuffer(input, output);
+    case IoctlCommand::IocFreeSpaceCommand:
+        return FreeSpace(input, output);
     default:
         break;
     }
@@ -91,6 +93,20 @@ u32 nvhost_as_gpu::AllocateSpace(const std::vector<u8>& input, std::vector<u8>&
     return result;
 }
 
+u32 nvhost_as_gpu::FreeSpace(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlFreeSpace params{};
+    std::memcpy(&params, input.data(), input.size());
+
+    LOG_DEBUG(Service_NVDRV, "called, offset={:X}, pages={:X}, page_size={:X}", params.offset,
+              params.pages, params.page_size);
+
+    system.GPU().MemoryManager().Unmap(params.offset,
+                                       static_cast<std::size_t>(params.pages) * params.page_size);
+
+    std::memcpy(output.data(), &params, output.size());
+    return NvErrCodes::Success;
+}
+
 u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) {
     const auto num_entries = input.size() / sizeof(IoctlRemapEntry);
 

src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h

@@ -82,6 +82,7 @@ private:
         IocBindChannelCommand = 0x40044101,
         IocGetVaRegionsCommand = 0xC0404108,
         IocUnmapBufferCommand = 0xC0084105,
+        IocFreeSpaceCommand = 0xC0104103,
     };
 
     struct IoctlInitalizeEx {
@@ -107,6 +108,13 @@ private:
     };
     static_assert(sizeof(IoctlAllocSpace) == 24, "IoctlInitalizeEx is incorrect size");
 
+    struct IoctlFreeSpace {
+        u64_le offset;
+        u32_le pages;
+        u32_le page_size;
+    };
+    static_assert(sizeof(IoctlFreeSpace) == 16, "IoctlFreeSpace is incorrect size");
+
     struct IoctlRemapEntry {
         u16_le flags;
         u16_le kind;
@@ -162,6 +170,7 @@ private:
     u32 Remap(const std::vector<u8>& input, std::vector<u8>& output);
     u32 MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output);
     u32 UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 FreeSpace(const std::vector<u8>& input, std::vector<u8>& output);
     u32 BindChannel(const std::vector<u8>& input, std::vector<u8>& output);
     u32 GetVARegions(const std::vector<u8>& input, std::vector<u8>& output);
 

src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp

@@ -15,8 +15,9 @@
 
 namespace Service::Nvidia::Devices {
 
-nvhost_ctrl::nvhost_ctrl(Core::System& system, EventInterface& events_interface)
-    : nvdevice(system), events_interface{events_interface} {}
+nvhost_ctrl::nvhost_ctrl(Core::System& system, EventInterface& events_interface,
+                         SyncpointManager& syncpoint_manager)
+    : nvdevice(system), events_interface{events_interface}, syncpoint_manager{syncpoint_manager} {}
 nvhost_ctrl::~nvhost_ctrl() = default;
 
 u32 nvhost_ctrl::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
@@ -36,8 +37,8 @@ u32 nvhost_ctrl::ioctl(Ioctl command, const std::vector<u8>& input, const std::v
         return IocCtrlEventRegister(input, output);
     case IoctlCommand::IocCtrlEventUnregisterCommand:
         return IocCtrlEventUnregister(input, output);
-    case IoctlCommand::IocCtrlEventSignalCommand:
-        return IocCtrlEventSignal(input, output);
+    case IoctlCommand::IocCtrlClearEventWaitCommand:
+        return IocCtrlClearEventWait(input, output);
     default:
         UNIMPLEMENTED_MSG("Unimplemented ioctl");
         return 0;
@@ -70,19 +71,33 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>&
         return NvResult::BadParameter;
     }
 
+    if (syncpoint_manager.IsSyncpointExpired(params.syncpt_id, params.threshold)) {
+        params.value = syncpoint_manager.GetSyncpointMin(params.syncpt_id);
+        std::memcpy(output.data(), &params, sizeof(params));
+        return NvResult::Success;
+    }
+
+    if (const auto new_value = syncpoint_manager.RefreshSyncpoint(params.syncpt_id);
+        syncpoint_manager.IsSyncpointExpired(params.syncpt_id, params.threshold)) {
+        params.value = new_value;
+        std::memcpy(output.data(), &params, sizeof(params));
+        return NvResult::Success;
+    }
+
     auto event = events_interface.events[event_id];
     auto& gpu = system.GPU();
+
     // This is mostly to take into account unimplemented features. As synced
     // gpu is always synced.
     if (!gpu.IsAsync()) {
-        event.writable->Signal();
+        event.event.writable->Signal();
         return NvResult::Success;
     }
     auto lock = gpu.LockSync();
-    const u32 current_syncpoint_value = gpu.GetSyncpointValue(params.syncpt_id);
+    const u32 current_syncpoint_value = event.fence.value;
     const s32 diff = current_syncpoint_value - params.threshold;
     if (diff >= 0) {
-        event.writable->Signal();
+        event.event.writable->Signal();
         params.value = current_syncpoint_value;
         std::memcpy(output.data(), &params, sizeof(params));
         return NvResult::Success;
@@ -109,7 +124,7 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>&
             params.value = ((params.syncpt_id & 0xfff) << 16) | 0x10000000;
         }
         params.value |= event_id;
-        event.writable->Clear();
+        event.event.writable->Clear();
         gpu.RegisterSyncptInterrupt(params.syncpt_id, target_value);
         if (!is_async && ctrl.fresh_call) {
             ctrl.must_delay = true;
@@ -154,24 +169,22 @@ u32 nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input, std::vecto
     return NvResult::Success;
 }
 
-u32 nvhost_ctrl::IocCtrlEventSignal(const std::vector<u8>& input, std::vector<u8>& output) {
+u32 nvhost_ctrl::IocCtrlClearEventWait(const std::vector<u8>& input, std::vector<u8>& output) {
     IocCtrlEventSignalParams params{};
     std::memcpy(&params, input.data(), sizeof(params));
-    // TODO(Blinkhawk): This is normally called when an NvEvents timeout on WaitSynchronization
-    // It is believed from RE to cancel the GPU Event. However, better research is required
-    u32 event_id = params.user_event_id & 0x00FF;
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called, user_event_id: {:X}", event_id);
+
+    u32 event_id = params.event_id & 0x00FF;
+    LOG_WARNING(Service_NVDRV, "cleared event wait on, event_id: {:X}", event_id);
+
     if (event_id >= MaxNvEvents) {
         return NvResult::BadParameter;
     }
     if (events_interface.status[event_id] == EventState::Waiting) {
-        auto& gpu = system.GPU();
-        if (gpu.CancelSyncptInterrupt(events_interface.assigned_syncpt[event_id],
-                                      events_interface.assigned_value[event_id])) {
-            events_interface.LiberateEvent(event_id);
-            events_interface.events[event_id].writable->Signal();
-        }
+        events_interface.LiberateEvent(event_id);
     }
+
+    syncpoint_manager.RefreshSyncpoint(events_interface.events[event_id].fence.id);
+
     return NvResult::Success;
 }
 

src/core/hle/service/nvdrv/devices/nvhost_ctrl.h

@@ -14,7 +14,8 @@ namespace Service::Nvidia::Devices {
 
 class nvhost_ctrl final : public nvdevice {
 public:
-    explicit nvhost_ctrl(Core::System& system, EventInterface& events_interface);
+    explicit nvhost_ctrl(Core::System& system, EventInterface& events_interface,
+                         SyncpointManager& syncpoint_manager);
     ~nvhost_ctrl() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
@@ -31,7 +32,7 @@ private:
         IocSyncptWaitexCommand = 0xC0100019,
         IocSyncptReadMaxCommand = 0xC008001A,
         IocGetConfigCommand = 0xC183001B,
-        IocCtrlEventSignalCommand = 0xC004001C,
+        IocCtrlClearEventWaitCommand = 0xC004001C,
         IocCtrlEventWaitCommand = 0xC010001D,
         IocCtrlEventWaitAsyncCommand = 0xC010001E,
         IocCtrlEventRegisterCommand = 0xC004001F,
@@ -94,7 +95,7 @@ private:
     static_assert(sizeof(IocGetConfigParams) == 387, "IocGetConfigParams is incorrect size");
 
     struct IocCtrlEventSignalParams {
-        u32_le user_event_id;
+        u32_le event_id;
     };
     static_assert(sizeof(IocCtrlEventSignalParams) == 4,
                   "IocCtrlEventSignalParams is incorrect size");
@@ -142,9 +143,10 @@ private:
 
     u32 IocCtrlEventUnregister(const std::vector<u8>& input, std::vector<u8>& output);
 
-    u32 IocCtrlEventSignal(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 IocCtrlClearEventWait(const std::vector<u8>& input, std::vector<u8>& output);
 
     EventInterface& events_interface;
+    SyncpointManager& syncpoint_manager;
 };
 
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp

@@ -7,14 +7,20 @@
 #include "common/logging/log.h"
 #include "core/core.h"
 #include "core/hle/service/nvdrv/devices/nvhost_gpu.h"
+#include "core/hle/service/nvdrv/syncpoint_manager.h"
 #include "core/memory.h"
 #include "video_core/gpu.h"
 #include "video_core/memory_manager.h"
 
 namespace Service::Nvidia::Devices {
 
-nvhost_gpu::nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev)
-    : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {}
+nvhost_gpu::nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev,
+                       SyncpointManager& syncpoint_manager)
+    : nvdevice(system), nvmap_dev(std::move(nvmap_dev)), syncpoint_manager{syncpoint_manager} {
+    channel_fence.id = syncpoint_manager.AllocateSyncpoint();
+    channel_fence.value = system.GPU().GetSyncpointValue(channel_fence.id);
+}
+
 nvhost_gpu::~nvhost_gpu() = default;
 
 u32 nvhost_gpu::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
@@ -126,10 +132,10 @@ u32 nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& ou
                 params.num_entries, params.flags, params.unk0, params.unk1, params.unk2,
                 params.unk3);
 
-    auto& gpu = system.GPU();
-    params.fence_out.id = assigned_syncpoints;
-    params.fence_out.value = gpu.GetSyncpointValue(assigned_syncpoints);
-    assigned_syncpoints++;
+    channel_fence.value = system.GPU().GetSyncpointValue(channel_fence.id);
+
+    params.fence_out = channel_fence;
+
     std::memcpy(output.data(), &params, output.size());
     return 0;
 }
@@ -145,37 +151,98 @@ u32 nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector<
     return 0;
 }
 
+static std::vector<Tegra::CommandHeader> BuildWaitCommandList(Fence fence) {
+    return {
+        Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceValue, 1,
+                                  Tegra::SubmissionMode::Increasing),
+        {fence.value},
+        Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceAction, 1,
+                                  Tegra::SubmissionMode::Increasing),
+        Tegra::GPU::FenceAction::Build(Tegra::GPU::FenceOperation::Acquire, fence.id),
+    };
+}
+
+static std::vector<Tegra::CommandHeader> BuildIncrementCommandList(Fence fence, u32 add_increment) {
+    std::vector<Tegra::CommandHeader> result{
+        Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceValue, 1,
+                                  Tegra::SubmissionMode::Increasing),
+        {}};
+
+    for (u32 count = 0; count < add_increment; ++count) {
+        result.emplace_back(Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceAction, 1,
+                                                      Tegra::SubmissionMode::Increasing));
+        result.emplace_back(
+            Tegra::GPU::FenceAction::Build(Tegra::GPU::FenceOperation::Increment, fence.id));
+    }
+
+    return result;
+}
+
+static std::vector<Tegra::CommandHeader> BuildIncrementWithWfiCommandList(Fence fence,
+                                                                          u32 add_increment) {
+    std::vector<Tegra::CommandHeader> result{
+        Tegra::BuildCommandHeader(Tegra::BufferMethods::WaitForInterrupt, 1,
+                                  Tegra::SubmissionMode::Increasing),
+        {}};
+    const std::vector<Tegra::CommandHeader> increment{
+        BuildIncrementCommandList(fence, add_increment)};
+
+    result.insert(result.end(), increment.begin(), increment.end());
+
+    return result;
+}
+
+u32 nvhost_gpu::SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector<u8>& output,
+                                 Tegra::CommandList&& entries) {
+    LOG_TRACE(Service_NVDRV, "called, gpfifo={:X}, num_entries={:X}, flags={:X}", params.address,
+              params.num_entries, params.flags.raw);
+
+    auto& gpu = system.GPU();
+
+    params.fence_out.id = channel_fence.id;
+
+    if (params.flags.add_wait.Value() &&
+        !syncpoint_manager.IsSyncpointExpired(params.fence_out.id, params.fence_out.value)) {
+        gpu.PushGPUEntries(Tegra::CommandList{BuildWaitCommandList(params.fence_out)});
+    }
+
+    if (params.flags.add_increment.Value() || params.flags.increment.Value()) {
+        const u32 increment_value = params.flags.increment.Value() ? params.fence_out.value : 0;
+        params.fence_out.value = syncpoint_manager.IncreaseSyncpoint(
+            params.fence_out.id, params.AddIncrementValue() + increment_value);
+    } else {
+        params.fence_out.value = syncpoint_manager.GetSyncpointMax(params.fence_out.id);
+    }
+
+    entries.RefreshIntegrityChecks(gpu);
+    gpu.PushGPUEntries(std::move(entries));
+
+    if (params.flags.add_increment.Value()) {
+        if (params.flags.suppress_wfi) {
+            gpu.PushGPUEntries(Tegra::CommandList{
+                BuildIncrementCommandList(params.fence_out, params.AddIncrementValue())});
+        } else {
+            gpu.PushGPUEntries(Tegra::CommandList{
+                BuildIncrementWithWfiCommandList(params.fence_out, params.AddIncrementValue())});
+        }
+    }
+
+    std::memcpy(output.data(), &params, sizeof(IoctlSubmitGpfifo));
+    return 0;
+}
+
 u32 nvhost_gpu::SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& output) {
     if (input.size() < sizeof(IoctlSubmitGpfifo)) {
         UNIMPLEMENTED();
     }
     IoctlSubmitGpfifo params{};
     std::memcpy(&params, input.data(), sizeof(IoctlSubmitGpfifo));
-    LOG_TRACE(Service_NVDRV, "called, gpfifo={:X}, num_entries={:X}, flags={:X}", params.address,
-              params.num_entries, params.flags.raw);
-
-    ASSERT_MSG(input.size() == sizeof(IoctlSubmitGpfifo) +
-                                   params.num_entries * sizeof(Tegra::CommandListHeader),
-               "Incorrect input size");
 
     Tegra::CommandList entries(params.num_entries);
-    std::memcpy(entries.data(), &input[sizeof(IoctlSubmitGpfifo)],
+    std::memcpy(entries.command_lists.data(), &input[sizeof(IoctlSubmitGpfifo)],
                 params.num_entries * sizeof(Tegra::CommandListHeader));
 
-    UNIMPLEMENTED_IF(params.flags.add_wait.Value() != 0);
-    UNIMPLEMENTED_IF(params.flags.add_increment.Value() != 0);
-
-    auto& gpu = system.GPU();
-    u32 current_syncpoint_value = gpu.GetSyncpointValue(params.fence_out.id);
-    if (params.flags.increment.Value()) {
-        params.fence_out.value += current_syncpoint_value;
-    } else {
-        params.fence_out.value = current_syncpoint_value;
-    }
-    gpu.PushGPUEntries(std::move(entries));
-
-    std::memcpy(output.data(), &params, sizeof(IoctlSubmitGpfifo));
-    return 0;
+    return SubmitGPFIFOImpl(params, output, std::move(entries));
 }
 
 u32 nvhost_gpu::KickoffPB(const std::vector<u8>& input, std::vector<u8>& output,
@@ -185,31 +252,17 @@ u32 nvhost_gpu::KickoffPB(const std::vector<u8>& input, std::vector<u8>& output,
     }
     IoctlSubmitGpfifo params{};
     std::memcpy(&params, input.data(), sizeof(IoctlSubmitGpfifo));
-    LOG_TRACE(Service_NVDRV, "called, gpfifo={:X}, num_entries={:X}, flags={:X}", params.address,
-              params.num_entries, params.flags.raw);
 
     Tegra::CommandList entries(params.num_entries);
     if (version == IoctlVersion::Version2) {
-        std::memcpy(entries.data(), input2.data(),
+        std::memcpy(entries.command_lists.data(), input2.data(),
                     params.num_entries * sizeof(Tegra::CommandListHeader));
     } else {
-        system.Memory().ReadBlock(params.address, entries.data(),
+        system.Memory().ReadBlock(params.address, entries.command_lists.data(),
                                   params.num_entries * sizeof(Tegra::CommandListHeader));
     }
-    UNIMPLEMENTED_IF(params.flags.add_wait.Value() != 0);
-    UNIMPLEMENTED_IF(params.flags.add_increment.Value() != 0);
 
-    auto& gpu = system.GPU();
-    u32 current_syncpoint_value = gpu.GetSyncpointValue(params.fence_out.id);
-    if (params.flags.increment.Value()) {
-        params.fence_out.value += current_syncpoint_value;
-    } else {
-        params.fence_out.value = current_syncpoint_value;
-    }
-    gpu.PushGPUEntries(std::move(entries));
-
-    std::memcpy(output.data(), &params, output.size());
-    return 0;
+    return SubmitGPFIFOImpl(params, output, std::move(entries));
 }
 
 u32 nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) {

src/core/hle/service/nvdrv/devices/nvhost_gpu.h

@@ -11,6 +11,11 @@
 #include "common/swap.h"
 #include "core/hle/service/nvdrv/devices/nvdevice.h"
 #include "core/hle/service/nvdrv/nvdata.h"
+#include "video_core/dma_pusher.h"
+
+namespace Service::Nvidia {
+class SyncpointManager;
+}
 
 namespace Service::Nvidia::Devices {
 
@@ -21,7 +26,8 @@ constexpr u32 NVGPU_IOCTL_CHANNEL_KICKOFF_PB(0x1b);
 
 class nvhost_gpu final : public nvdevice {
 public:
-    explicit nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev);
+    explicit nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev,
+                        SyncpointManager& syncpoint_manager);
     ~nvhost_gpu() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
@@ -162,10 +168,15 @@ private:
             u32_le raw;
             BitField<0, 1, u32_le> add_wait;      // append a wait sync_point to the list
             BitField<1, 1, u32_le> add_increment; // append an increment to the list
-            BitField<2, 1, u32_le> new_hw_format; // Mostly ignored
+            BitField<2, 1, u32_le> new_hw_format; // mostly ignored
+            BitField<4, 1, u32_le> suppress_wfi;  // suppress wait for interrupt
             BitField<8, 1, u32_le> increment;     // increment the returned fence
         } flags;
         Fence fence_out; // returned new fence object for others to wait on
+
+        u32 AddIncrementValue() const {
+            return flags.add_increment.Value() << 1;
+        }
     };
     static_assert(sizeof(IoctlSubmitGpfifo) == 16 + sizeof(Fence),
                   "IoctlSubmitGpfifo is incorrect size");
@@ -190,6 +201,8 @@ private:
     u32 SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output);
     u32 AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output);
     u32 AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector<u8>& output,
+                         Tegra::CommandList&& entries);
     u32 SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& output);
     u32 KickoffPB(const std::vector<u8>& input, std::vector<u8>& output,
                   const std::vector<u8>& input2, IoctlVersion version);
@@ -198,7 +211,8 @@ private:
     u32 ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output);
 
     std::shared_ptr<nvmap> nvmap_dev;
-    u32 assigned_syncpoints{};
+    SyncpointManager& syncpoint_manager;
+    Fence channel_fence;
 };
 
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp

@@ -2,15 +2,17 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <cstring>
-
 #include "common/assert.h"
 #include "common/logging/log.h"
+#include "core/core.h"
 #include "core/hle/service/nvdrv/devices/nvhost_nvdec.h"
+#include "video_core/memory_manager.h"
+#include "video_core/renderer_base.h"
 
 namespace Service::Nvidia::Devices {
 
-nvhost_nvdec::nvhost_nvdec(Core::System& system) : nvdevice(system) {}
+nvhost_nvdec::nvhost_nvdec(Core::System& system, std::shared_ptr<nvmap> nvmap_dev)
+    : nvhost_nvdec_common(system, std::move(nvmap_dev)) {}
 nvhost_nvdec::~nvhost_nvdec() = default;
 
 u32 nvhost_nvdec::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
@@ -21,7 +23,7 @@ u32 nvhost_nvdec::ioctl(Ioctl command, const std::vector<u8>& input, const std::
 
     switch (static_cast<IoctlCommand>(command.raw)) {
     case IoctlCommand::IocSetNVMAPfdCommand:
-        return SetNVMAPfd(input, output);
+        return SetNVMAPfd(input);
     case IoctlCommand::IocSubmit:
         return Submit(input, output);
     case IoctlCommand::IocGetSyncpoint:
@@ -29,79 +31,29 @@ u32 nvhost_nvdec::ioctl(Ioctl command, const std::vector<u8>& input, const std::
     case IoctlCommand::IocGetWaitbase:
         return GetWaitbase(input, output);
     case IoctlCommand::IocMapBuffer:
-        return MapBuffer(input, output);
+    case IoctlCommand::IocMapBuffer2:
+    case IoctlCommand::IocMapBuffer3:
     case IoctlCommand::IocMapBufferEx:
-        return MapBufferEx(input, output);
-    case IoctlCommand::IocUnmapBufferEx:
-        return UnmapBufferEx(input, output);
+        return MapBuffer(input, output);
+    case IoctlCommand::IocUnmapBufferEx: {
+        // This command is sent when the video stream has ended, flush all video contexts
+        // This is usually sent in the folowing order: vic, nvdec, vic.
+        // Inform the GPU to clear any remaining nvdec buffers when this is detected.
+        LOG_INFO(Service_NVDRV, "NVDEC video stream ended");
+        Tegra::ChCommandHeaderList cmdlist(1);
+        cmdlist[0] = Tegra::ChCommandHeader{0xDEADB33F};
+        system.GPU().PushCommandBuffer(cmdlist);
+        [[fallthrough]]; // fallthrough to unmap buffers
+    };
+    case IoctlCommand::IocUnmapBuffer:
+    case IoctlCommand::IocUnmapBuffer2:
+    case IoctlCommand::IocUnmapBuffer3:
+        return UnmapBuffer(input, output);
+    case IoctlCommand::IocSetSubmitTimeout:
+        return SetSubmitTimeout(input, output);
     }
 
-    UNIMPLEMENTED_MSG("Unimplemented ioctl");
-    return 0;
-}
-
-u32 nvhost_nvdec::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlSetNvmapFD params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlSetNvmapFD));
-    LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd);
-
-    nvmap_fd = params.nvmap_fd;
-    return 0;
-}
-
-u32 nvhost_nvdec::Submit(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlSubmit params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlSubmit));
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called");
-    std::memcpy(output.data(), &params, sizeof(IoctlSubmit));
-    return 0;
-}
-
-u32 nvhost_nvdec::GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlGetSyncpoint params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlGetSyncpoint));
-    LOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown);
-    params.value = 0; // Seems to be hard coded at 0
-    std::memcpy(output.data(), &params, sizeof(IoctlGetSyncpoint));
-    return 0;
-}
-
-u32 nvhost_nvdec::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlGetWaitbase params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlGetWaitbase));
-    LOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown);
-    params.value = 0; // Seems to be hard coded at 0
-    std::memcpy(output.data(), &params, sizeof(IoctlGetWaitbase));
-    return 0;
-}
-
-u32 nvhost_nvdec::MapBuffer(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlMapBuffer params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlMapBuffer));
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called with address={:08X}{:08X}", params.address_2,
-                params.address_1);
-    params.address_1 = 0;
-    params.address_2 = 0;
-    std::memcpy(output.data(), &params, sizeof(IoctlMapBuffer));
-    return 0;
-}
-
-u32 nvhost_nvdec::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlMapBufferEx params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlMapBufferEx));
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called with address={:08X}{:08X}", params.address_2,
-                params.address_1);
-    params.address_1 = 0;
-    params.address_2 = 0;
-    std::memcpy(output.data(), &params, sizeof(IoctlMapBufferEx));
-    return 0;
-}
-
-u32 nvhost_nvdec::UnmapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlUnmapBufferEx params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlUnmapBufferEx));
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called");
-    std::memcpy(output.data(), &params, sizeof(IoctlUnmapBufferEx));
+    UNIMPLEMENTED_MSG("Unimplemented ioctl 0x{:X}", command.raw);
     return 0;
 }
 

src/core/hle/service/nvdrv/devices/nvhost_nvdec.h

@@ -4,16 +4,14 @@
 
 #pragma once
 
-#include <vector>
-#include "common/common_types.h"
-#include "common/swap.h"
-#include "core/hle/service/nvdrv/devices/nvdevice.h"
+#include <memory>
+#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h"
 
 namespace Service::Nvidia::Devices {
 
-class nvhost_nvdec final : public nvdevice {
+class nvhost_nvdec final : public nvhost_nvdec_common {
 public:
-    explicit nvhost_nvdec(Core::System& system);
+    explicit nvhost_nvdec(Core::System& system, std::shared_ptr<nvmap> nvmap_dev);
     ~nvhost_nvdec() override;
 
     u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
@@ -27,62 +25,15 @@ private:
         IocGetSyncpoint = 0xC0080002,
         IocGetWaitbase = 0xC0080003,
         IocMapBuffer = 0xC01C0009,
+        IocMapBuffer2 = 0xC16C0009,
+        IocMapBuffer3 = 0xC15C0009,
         IocMapBufferEx = 0xC0A40009,
-        IocUnmapBufferEx = 0xC0A4000A,
+        IocUnmapBuffer = 0xC0A4000A,
+        IocUnmapBuffer2 = 0xC16C000A,
+        IocUnmapBufferEx = 0xC01C000A,
+        IocUnmapBuffer3 = 0xC15C000A,
+        IocSetSubmitTimeout = 0x40040007,
     };
-
-    struct IoctlSetNvmapFD {
-        u32_le nvmap_fd;
-    };
-    static_assert(sizeof(IoctlSetNvmapFD) == 0x4, "IoctlSetNvmapFD is incorrect size");
-
-    struct IoctlSubmit {
-        INSERT_PADDING_BYTES(0x40); // TODO(DarkLordZach): RE this structure
-    };
-    static_assert(sizeof(IoctlSubmit) == 0x40, "IoctlSubmit has incorrect size");
-
-    struct IoctlGetSyncpoint {
-        u32 unknown; // seems to be ignored? Nintendo added this
-        u32 value;
-    };
-    static_assert(sizeof(IoctlGetSyncpoint) == 0x08, "IoctlGetSyncpoint has incorrect size");
-
-    struct IoctlGetWaitbase {
-        u32 unknown; // seems to be ignored? Nintendo added this
-        u32 value;
-    };
-    static_assert(sizeof(IoctlGetWaitbase) == 0x08, "IoctlGetWaitbase has incorrect size");
-
-    struct IoctlMapBuffer {
-        u32 unknown;
-        u32 address_1;
-        u32 address_2;
-        INSERT_PADDING_BYTES(0x10); // TODO(DarkLordZach): RE this structure
-    };
-    static_assert(sizeof(IoctlMapBuffer) == 0x1C, "IoctlMapBuffer is incorrect size");
-
-    struct IoctlMapBufferEx {
-        u32 unknown;
-        u32 address_1;
-        u32 address_2;
-        INSERT_PADDING_BYTES(0x98); // TODO(DarkLordZach): RE this structure
-    };
-    static_assert(sizeof(IoctlMapBufferEx) == 0xA4, "IoctlMapBufferEx has incorrect size");
-
-    struct IoctlUnmapBufferEx {
-        INSERT_PADDING_BYTES(0xA4); // TODO(DarkLordZach): RE this structure
-    };
-    static_assert(sizeof(IoctlUnmapBufferEx) == 0xA4, "IoctlUnmapBufferEx has incorrect size");
-
-    u32_le nvmap_fd{};
-
-    u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 Submit(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 MapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 UnmapBufferEx(const std::vector<u8>& input, std::vector<u8>& output);
 };
 
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp

@@ -0,0 +1,234 @@
+// Copyright 2020 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <algorithm>
+#include <cstring>
+
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "common/logging/log.h"
+#include "core/core.h"
+#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h"
+#include "core/hle/service/nvdrv/devices/nvmap.h"
+#include "core/memory.h"
+#include "video_core/memory_manager.h"
+#include "video_core/renderer_base.h"
+
+namespace Service::Nvidia::Devices {
+
+namespace {
+// Splice vectors will copy count amount of type T from the input vector into the dst vector.
+template <typename T>
+std::size_t SpliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count,
+                          std::size_t offset) {
+    std::memcpy(dst.data(), input.data() + offset, count * sizeof(T));
+    offset += count * sizeof(T);
+    return offset;
+}
+
+// Write vectors will write data to the output buffer
+template <typename T>
+std::size_t WriteVectors(std::vector<u8>& dst, const std::vector<T>& src, std::size_t offset) {
+    std::memcpy(dst.data() + offset, src.data(), src.size() * sizeof(T));
+    offset += src.size() * sizeof(T);
+    return offset;
+}
+} // Anonymous namespace
+
+namespace NvErrCodes {
+constexpr u32 Success{};
+[[maybe_unused]] constexpr u32 OutOfMemory{static_cast<u32>(-12)};
+constexpr u32 InvalidInput{static_cast<u32>(-22)};
+} // namespace NvErrCodes
+
+nvhost_nvdec_common::nvhost_nvdec_common(Core::System& system, std::shared_ptr<nvmap> nvmap_dev)
+    : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {}
+nvhost_nvdec_common::~nvhost_nvdec_common() = default;
+
+u32 nvhost_nvdec_common::SetNVMAPfd(const std::vector<u8>& input) {
+    IoctlSetNvmapFD params{};
+    std::memcpy(&params, input.data(), sizeof(IoctlSetNvmapFD));
+    LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd);
+
+    nvmap_fd = params.nvmap_fd;
+    return 0;
+}
+
+u32 nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlSubmit params{};
+    std::memcpy(&params, input.data(), sizeof(IoctlSubmit));
+    LOG_DEBUG(Service_NVDRV, "called NVDEC Submit, cmd_buffer_count={}", params.cmd_buffer_count);
+
+    // Instantiate param buffers
+    std::size_t offset = sizeof(IoctlSubmit);
+    std::vector<CommandBuffer> command_buffers(params.cmd_buffer_count);
+    std::vector<Reloc> relocs(params.relocation_count);
+    std::vector<u32> reloc_shifts(params.relocation_count);
+    std::vector<SyncptIncr> syncpt_increments(params.syncpoint_count);
+    std::vector<SyncptIncr> wait_checks(params.syncpoint_count);
+    std::vector<Fence> fences(params.fence_count);
+
+    // Splice input into their respective buffers
+    offset = SpliceVectors(input, command_buffers, params.cmd_buffer_count, offset);
+    offset = SpliceVectors(input, relocs, params.relocation_count, offset);
+    offset = SpliceVectors(input, reloc_shifts, params.relocation_count, offset);
+    offset = SpliceVectors(input, syncpt_increments, params.syncpoint_count, offset);
+    offset = SpliceVectors(input, wait_checks, params.syncpoint_count, offset);
+    offset = SpliceVectors(input, fences, params.fence_count, offset);
+
+    // TODO(ameerj): For async gpu, utilize fences for syncpoint 'max' increment
+
+    auto& gpu = system.GPU();
+
+    for (const auto& cmd_buffer : command_buffers) {
+        auto object = nvmap_dev->GetObject(cmd_buffer.memory_id);
+        ASSERT_OR_EXECUTE(object, return NvErrCodes::InvalidInput;);
+        const auto map = FindBufferMap(object->dma_map_addr);
+        if (!map) {
+            LOG_ERROR(Service_NVDRV, "Tried to submit an invalid offset 0x{:X} dma 0x{:X}",
+                      object->addr, object->dma_map_addr);
+            return 0;
+        }
+        Tegra::ChCommandHeaderList cmdlist(cmd_buffer.word_count);
+        gpu.MemoryManager().ReadBlock(map->StartAddr() + cmd_buffer.offset, cmdlist.data(),
+                                      cmdlist.size() * sizeof(u32));
+        gpu.PushCommandBuffer(cmdlist);
+    }
+
+    std::memcpy(output.data(), &params, sizeof(IoctlSubmit));
+    // Some games expect command_buffers to be written back
+    offset = sizeof(IoctlSubmit);
+    offset = WriteVectors(output, command_buffers, offset);
+    offset = WriteVectors(output, relocs, offset);
+    offset = WriteVectors(output, reloc_shifts, offset);
+    offset = WriteVectors(output, syncpt_increments, offset);
+    offset = WriteVectors(output, wait_checks, offset);
+
+    return NvErrCodes::Success;
+}
+
+u32 nvhost_nvdec_common::GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlGetSyncpoint params{};
+    std::memcpy(&params, input.data(), sizeof(IoctlGetSyncpoint));
+    LOG_DEBUG(Service_NVDRV, "called GetSyncpoint, id={}", params.param);
+
+    // We found that implementing this causes deadlocks with async gpu, along with degraded
+    // performance. TODO: RE the nvdec async implementation
+    params.value = 0;
+    std::memcpy(output.data(), &params, sizeof(IoctlGetSyncpoint));
+
+    return NvErrCodes::Success;
+}
+
+u32 nvhost_nvdec_common::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlGetWaitbase params{};
+    std::memcpy(&params, input.data(), sizeof(IoctlGetWaitbase));
+    params.value = 0; // Seems to be hard coded at 0
+    std::memcpy(output.data(), &params, sizeof(IoctlGetWaitbase));
+    return 0;
+}
+
+u32 nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlMapBuffer params{};
+    std::memcpy(&params, input.data(), sizeof(IoctlMapBuffer));
+    std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries);
+
+    SpliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer));
+
+    auto& gpu = system.GPU();
+
+    for (auto& cmf_buff : cmd_buffer_handles) {
+        auto object{nvmap_dev->GetObject(cmf_buff.map_handle)};
+        if (!object) {
+            LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle);
+            std::memcpy(output.data(), &params, output.size());
+            return NvErrCodes::InvalidInput;
+        }
+        if (object->dma_map_addr == 0) {
+            // NVDEC and VIC memory is in the 32-bit address space
+            // MapAllocate32 will attempt to map a lower 32-bit value in the shared gpu memory space
+            const GPUVAddr low_addr = gpu.MemoryManager().MapAllocate32(object->addr, object->size);
+            object->dma_map_addr = static_cast<u32>(low_addr);
+            // Ensure that the dma_map_addr is indeed in the lower 32-bit address space.
+            ASSERT(object->dma_map_addr == low_addr);
+        }
+        if (!object->dma_map_addr) {
+            LOG_ERROR(Service_NVDRV, "failed to map size={}", object->size);
+        } else {
+            cmf_buff.map_address = object->dma_map_addr;
+            AddBufferMap(object->dma_map_addr, object->size, object->addr,
+                         object->status == nvmap::Object::Status::Allocated);
+        }
+    }
+    std::memcpy(output.data(), &params, sizeof(IoctlMapBuffer));
+    std::memcpy(output.data() + sizeof(IoctlMapBuffer), cmd_buffer_handles.data(),
+                cmd_buffer_handles.size() * sizeof(MapBufferEntry));
+
+    return NvErrCodes::Success;
+}
+
+u32 nvhost_nvdec_common::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) {
+    IoctlMapBuffer params{};
+    std::memcpy(&params, input.data(), sizeof(IoctlMapBuffer));
+    std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries);
+    SpliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer));
+
+    auto& gpu = system.GPU();
+
+    for (auto& cmf_buff : cmd_buffer_handles) {
+        const auto object{nvmap_dev->GetObject(cmf_buff.map_handle)};
+        if (!object) {
+            LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle);
+            std::memcpy(output.data(), &params, output.size());
+            return NvErrCodes::InvalidInput;
+        }
+        if (const auto size{RemoveBufferMap(object->dma_map_addr)}; size) {
+            gpu.MemoryManager().Unmap(object->dma_map_addr, *size);
+        } else {
+            // This occurs quite frequently, however does not seem to impact functionality
+            LOG_DEBUG(Service_NVDRV, "invalid offset=0x{:X} dma=0x{:X}", object->addr,
+                      object->dma_map_addr);
+        }
+        object->dma_map_addr = 0;
+    }
+    std::memset(output.data(), 0, output.size());
+    return NvErrCodes::Success;
+}
+
+u32 nvhost_nvdec_common::SetSubmitTimeout(const std::vector<u8>& input, std::vector<u8>& output) {
+    std::memcpy(&submit_timeout, input.data(), input.size());
+    LOG_WARNING(Service_NVDRV, "(STUBBED) called");
+    return NvErrCodes::Success;
+}
+
+std::optional<nvhost_nvdec_common::BufferMap> nvhost_nvdec_common::FindBufferMap(
+    GPUVAddr gpu_addr) const {
+    const auto it = std::find_if(
+        buffer_mappings.begin(), buffer_mappings.upper_bound(gpu_addr), [&](const auto& entry) {
+            return (gpu_addr >= entry.second.StartAddr() && gpu_addr < entry.second.EndAddr());
+        });
+
+    ASSERT(it != buffer_mappings.end());
+    return it->second;
+}
+
+void nvhost_nvdec_common::AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr,
+                                       bool is_allocated) {
+    buffer_mappings.insert_or_assign(gpu_addr, BufferMap{gpu_addr, size, cpu_addr, is_allocated});
+}
+
+std::optional<std::size_t> nvhost_nvdec_common::RemoveBufferMap(GPUVAddr gpu_addr) {
+    const auto iter{buffer_mappings.find(gpu_addr)};
+    if (iter == buffer_mappings.end()) {
+        return std::nullopt;
+    }
+    std::size_t size = 0;
+    if (iter->second.IsAllocated()) {
+        size = iter->second.Size();
+    }
+    buffer_mappings.erase(iter);
+    return size;
+}
+
+} // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h

@@ -0,0 +1,168 @@
+// Copyright 2020 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <map>
+#include <vector>
+#include "common/common_types.h"
+#include "common/swap.h"
+#include "core/hle/service/nvdrv/devices/nvdevice.h"
+
+namespace Service::Nvidia::Devices {
+class nvmap;
+
+class nvhost_nvdec_common : public nvdevice {
+public:
+    explicit nvhost_nvdec_common(Core::System& system, std::shared_ptr<nvmap> nvmap_dev);
+    ~nvhost_nvdec_common() override;
+
+    virtual u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
+                      std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl,
+                      IoctlVersion version) = 0;
+
+protected:
+    class BufferMap final {
+    public:
+        constexpr BufferMap() = default;
+
+        constexpr BufferMap(GPUVAddr start_addr, std::size_t size)
+            : start_addr{start_addr}, end_addr{start_addr + size} {}
+
+        constexpr BufferMap(GPUVAddr start_addr, std::size_t size, VAddr cpu_addr,
+                            bool is_allocated)
+            : start_addr{start_addr}, end_addr{start_addr + size}, cpu_addr{cpu_addr},
+              is_allocated{is_allocated} {}
+
+        constexpr VAddr StartAddr() const {
+            return start_addr;
+        }
+
+        constexpr VAddr EndAddr() const {
+            return end_addr;
+        }
+
+        constexpr std::size_t Size() const {
+            return end_addr - start_addr;
+        }
+
+        constexpr VAddr CpuAddr() const {
+            return cpu_addr;
+        }
+
+        constexpr bool IsAllocated() const {
+            return is_allocated;
+        }
+
+    private:
+        GPUVAddr start_addr{};
+        GPUVAddr end_addr{};
+        VAddr cpu_addr{};
+        bool is_allocated{};
+    };
+
+    struct IoctlSetNvmapFD {
+        u32_le nvmap_fd;
+    };
+    static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size");
+
+    struct IoctlSubmitCommandBuffer {
+        u32_le id;
+        u32_le offset;
+        u32_le count;
+    };
+    static_assert(sizeof(IoctlSubmitCommandBuffer) == 0xC,
+                  "IoctlSubmitCommandBuffer is incorrect size");
+    struct IoctlSubmit {
+        u32_le cmd_buffer_count;
+        u32_le relocation_count;
+        u32_le syncpoint_count;
+        u32_le fence_count;
+    };
+    static_assert(sizeof(IoctlSubmit) == 0x10, "IoctlSubmit has incorrect size");
+
+    struct CommandBuffer {
+        s32 memory_id;
+        u32 offset;
+        s32 word_count;
+    };
+    static_assert(sizeof(CommandBuffer) == 0xC, "CommandBuffer has incorrect size");
+
+    struct Reloc {
+        s32 cmdbuffer_memory;
+        s32 cmdbuffer_offset;
+        s32 target;
+        s32 target_offset;
+    };
+    static_assert(sizeof(Reloc) == 0x10, "CommandBuffer has incorrect size");
+
+    struct SyncptIncr {
+        u32 id;
+        u32 increments;
+    };
+    static_assert(sizeof(SyncptIncr) == 0x8, "CommandBuffer has incorrect size");
+
+    struct Fence {
+        u32 id;
+        u32 value;
+    };
+    static_assert(sizeof(Fence) == 0x8, "CommandBuffer has incorrect size");
+
+    struct IoctlGetSyncpoint {
+        // Input
+        u32_le param;
+        // Output
+        u32_le value;
+    };
+    static_assert(sizeof(IoctlGetSyncpoint) == 8, "IocGetIdParams has wrong size");
+
+    struct IoctlGetWaitbase {
+        u32_le unknown; // seems to be ignored? Nintendo added this
+        u32_le value;
+    };
+    static_assert(sizeof(IoctlGetWaitbase) == 0x8, "IoctlGetWaitbase is incorrect size");
+
+    struct IoctlMapBuffer {
+        u32_le num_entries;
+        u32_le data_address; // Ignored by the driver.
+        u32_le attach_host_ch_das;
+    };
+    static_assert(sizeof(IoctlMapBuffer) == 0x0C, "IoctlMapBuffer is incorrect size");
+
+    struct IocGetIdParams {
+        // Input
+        u32_le param;
+        // Output
+        u32_le value;
+    };
+    static_assert(sizeof(IocGetIdParams) == 8, "IocGetIdParams has wrong size");
+
+    // Used for mapping and unmapping command buffers
+    struct MapBufferEntry {
+        u32_le map_handle;
+        u32_le map_address;
+    };
+    static_assert(sizeof(IoctlMapBuffer) == 0x0C, "IoctlMapBuffer is incorrect size");
+
+    /// Ioctl command implementations
+    u32 SetNVMAPfd(const std::vector<u8>& input);
+    u32 Submit(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 MapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
+    u32 SetSubmitTimeout(const std::vector<u8>& input, std::vector<u8>& output);
+
+    std::optional<BufferMap> FindBufferMap(GPUVAddr gpu_addr) const;
+    void AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr, bool is_allocated);
+    std::optional<std::size_t> RemoveBufferMap(GPUVAddr gpu_addr);
+
+    u32_le nvmap_fd{};
+    u32_le submit_timeout{};
+    std::shared_ptr<nvmap> nvmap_dev;
+
+    // This is expected to be ordered, therefore we must use a map, not unordered_map
+    std::map<GPUVAddr, BufferMap> buffer_mappings;
+};
+}; // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvhost_vic.cpp

@@ -2,15 +2,17 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
-#include <cstring>
-
 #include "common/assert.h"
 #include "common/logging/log.h"
+#include "core/core.h"
 #include "core/hle/service/nvdrv/devices/nvhost_vic.h"
+#include "video_core/memory_manager.h"
+#include "video_core/renderer_base.h"
 
 namespace Service::Nvidia::Devices {
+nvhost_vic::nvhost_vic(Core::System& system, std::shared_ptr<nvmap> nvmap_dev)
+    : nvhost_nvdec_common(system, std::move(nvmap_dev)) {}
 
-nvhost_vic::nvhost_vic(Core::System& system) : nvdevice(system) {}
 nvhost_vic::~nvhost_vic() = default;
 
 u32 nvhost_vic::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
@@ -21,7 +23,7 @@ u32 nvhost_vic::ioctl(Ioctl command, const std::vector<u8>& input, const std::ve
 
     switch (static_cast<IoctlCommand>(command.raw)) {
     case IoctlCommand::IocSetNVMAPfdCommand:
-        return SetNVMAPfd(input, output);
+        return SetNVMAPfd(input);
     case IoctlCommand::IocSubmit:
         return Submit(input, output);
     case IoctlCommand::IocGetSyncpoint:
@@ -29,83 +31,19 @@ u32 nvhost_vic::ioctl(Ioctl command, const std::vector<u8>& input, const std::ve
     case IoctlCommand::IocGetWaitbase:
         return GetWaitbase(input, output);
     case IoctlCommand::IocMapBuffer:
-        return MapBuffer(input, output);
+    case IoctlCommand::IocMapBuffer2:
+    case IoctlCommand::IocMapBuffer3:
+    case IoctlCommand::IocMapBuffer4:
     case IoctlCommand::IocMapBufferEx:
         return MapBuffer(input, output);
+    case IoctlCommand::IocUnmapBuffer:
+    case IoctlCommand::IocUnmapBuffer2:
+    case IoctlCommand::IocUnmapBuffer3:
     case IoctlCommand::IocUnmapBufferEx:
-        return UnmapBufferEx(input, output);
+        return UnmapBuffer(input, output);
     }
 
-    UNIMPLEMENTED_MSG("Unimplemented ioctl");
-    return 0;
-}
-
-u32 nvhost_vic::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlSetNvmapFD params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlSetNvmapFD));
-    LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd);
-
-    nvmap_fd = params.nvmap_fd;
-    return 0;
-}
-
-u32 nvhost_vic::Submit(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlSubmit params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlSubmit));
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called");
-
-    // Workaround for Luigi's Mansion 3, as nvhost_vic is not implemented for asynch GPU
-    params.command_buffer = {};
-
-    std::memcpy(output.data(), &params, sizeof(IoctlSubmit));
-    return 0;
-}
-
-u32 nvhost_vic::GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlGetSyncpoint params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlGetSyncpoint));
-    LOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown);
-    params.value = 0; // Seems to be hard coded at 0
-    std::memcpy(output.data(), &params, sizeof(IoctlGetSyncpoint));
-    return 0;
-}
-
-u32 nvhost_vic::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlGetWaitbase params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlGetWaitbase));
-    LOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown);
-    params.value = 0; // Seems to be hard coded at 0
-    std::memcpy(output.data(), &params, sizeof(IoctlGetWaitbase));
-    return 0;
-}
-
-u32 nvhost_vic::MapBuffer(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlMapBuffer params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlMapBuffer));
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called with address={:08X}{:08X}", params.address_2,
-                params.address_1);
-    params.address_1 = 0;
-    params.address_2 = 0;
-    std::memcpy(output.data(), &params, sizeof(IoctlMapBuffer));
-    return 0;
-}
-
-u32 nvhost_vic::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlMapBufferEx params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlMapBufferEx));
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called with address={:08X}{:08X}", params.address_2,
-                params.address_1);
-    params.address_1 = 0;
-    params.address_2 = 0;
-    std::memcpy(output.data(), &params, sizeof(IoctlMapBufferEx));
-    return 0;
-}
-
-u32 nvhost_vic::UnmapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) {
-    IoctlUnmapBufferEx params{};
-    std::memcpy(&params, input.data(), sizeof(IoctlUnmapBufferEx));
-    LOG_WARNING(Service_NVDRV, "(STUBBED) called");
-    std::memcpy(output.data(), &params, sizeof(IoctlUnmapBufferEx));
+    UNIMPLEMENTED_MSG("Unimplemented ioctl 0x{:X}", command.raw);
     return 0;
 }
 

src/core/hle/service/nvdrv/devices/nvhost_vic.h

@@ -4,19 +4,15 @@
 
 #pragma once
 
-#include <array>
-#include <vector>
-#include "common/common_types.h"
-#include "common/swap.h"
-#include "core/hle/service/nvdrv/devices/nvdevice.h"
+#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h"
 
 namespace Service::Nvidia::Devices {
+class nvmap;
 
-class nvhost_vic final : public nvdevice {
+class nvhost_vic final : public nvhost_nvdec_common {
 public:
-    explicit nvhost_vic(Core::System& system);
-    ~nvhost_vic() override;
-
+    explicit nvhost_vic(Core::System& system, std::shared_ptr<nvmap> nvmap_dev);
+    ~nvhost_vic();
     u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2,
               std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl,
               IoctlVersion version) override;
@@ -28,74 +24,14 @@ private:
         IocGetSyncpoint = 0xC0080002,
         IocGetWaitbase = 0xC0080003,
         IocMapBuffer = 0xC01C0009,
+        IocMapBuffer2 = 0xC0340009,
+        IocMapBuffer3 = 0xC0140009,
+        IocMapBuffer4 = 0xC00C0009,
         IocMapBufferEx = 0xC03C0009,
-        IocUnmapBufferEx = 0xC03C000A,
+        IocUnmapBuffer = 0xC03C000A,
+        IocUnmapBuffer2 = 0xC034000A,
+        IocUnmapBuffer3 = 0xC00C000A,
+        IocUnmapBufferEx = 0xC01C000A,
     };
-
-    struct IoctlSetNvmapFD {
-        u32_le nvmap_fd;
-    };
-    static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size");
-
-    struct IoctlSubmitCommandBuffer {
-        u32 id;
-        u32 offset;
-        u32 count;
-    };
-    static_assert(sizeof(IoctlSubmitCommandBuffer) == 0xC,
-                  "IoctlSubmitCommandBuffer is incorrect size");
-
-    struct IoctlSubmit {
-        u32 command_buffer_count;
-        u32 relocations_count;
-        u32 syncpt_count;
-        u32 wait_count;
-        std::array<IoctlSubmitCommandBuffer, 4> command_buffer;
-    };
-    static_assert(sizeof(IoctlSubmit) == 0x40, "IoctlSubmit is incorrect size");
-
-    struct IoctlGetSyncpoint {
-        u32 unknown; // seems to be ignored? Nintendo added this
-        u32 value;
-    };
-    static_assert(sizeof(IoctlGetSyncpoint) == 0x8, "IoctlGetSyncpoint is incorrect size");
-
-    struct IoctlGetWaitbase {
-        u32 unknown; // seems to be ignored? Nintendo added this
-        u32 value;
-    };
-    static_assert(sizeof(IoctlGetWaitbase) == 0x8, "IoctlGetWaitbase is incorrect size");
-
-    struct IoctlMapBuffer {
-        u32 unknown;
-        u32 address_1;
-        u32 address_2;
-        INSERT_PADDING_BYTES(0x10); // TODO(DarkLordZach): RE this structure
-    };
-    static_assert(sizeof(IoctlMapBuffer) == 0x1C, "IoctlMapBuffer is incorrect size");
-
-    struct IoctlMapBufferEx {
-        u32 unknown;
-        u32 address_1;
-        u32 address_2;
-        INSERT_PADDING_BYTES(0x30); // TODO(DarkLordZach): RE this structure
-    };
-    static_assert(sizeof(IoctlMapBufferEx) == 0x3C, "IoctlMapBufferEx is incorrect size");
-
-    struct IoctlUnmapBufferEx {
-        INSERT_PADDING_BYTES(0x3C); // TODO(DarkLordZach): RE this structure
-    };
-    static_assert(sizeof(IoctlUnmapBufferEx) == 0x3C, "IoctlUnmapBufferEx is incorrect size");
-
-    u32_le nvmap_fd{};
-
-    u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 Submit(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 MapBuffer(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output);
-    u32 UnmapBufferEx(const std::vector<u8>& input, std::vector<u8>& output);
 };
-
 } // namespace Service::Nvidia::Devices

src/core/hle/service/nvdrv/devices/nvmap.h

@@ -37,6 +37,7 @@ public:
         VAddr addr;
         Status status;
         u32 refcount;
+        u32 dma_map_addr;
     };
 
     std::shared_ptr<Object> GetObject(u32 handle) const {

src/core/hle/service/nvdrv/nvdrv.cpp

@@ -21,6 +21,7 @@
 #include "core/hle/service/nvdrv/interface.h"
 #include "core/hle/service/nvdrv/nvdrv.h"
 #include "core/hle/service/nvdrv/nvmemp.h"
+#include "core/hle/service/nvdrv/syncpoint_manager.h"
 #include "core/hle/service/nvflinger/nvflinger.h"
 
 namespace Service::Nvidia {
@@ -36,24 +37,26 @@ void InstallInterfaces(SM::ServiceManager& service_manager, NVFlinger::NVFlinger
     nvflinger.SetNVDrvInstance(module_);
 }
 
-Module::Module(Core::System& system) {
+Module::Module(Core::System& system) : syncpoint_manager{system.GPU()} {
     auto& kernel = system.Kernel();
     for (u32 i = 0; i < MaxNvEvents; i++) {
         std::string event_label = fmt::format("NVDRV::NvEvent_{}", i);
-        events_interface.events[i] = Kernel::WritableEvent::CreateEventPair(kernel, event_label);
+        events_interface.events[i] = {Kernel::WritableEvent::CreateEventPair(kernel, event_label)};
         events_interface.status[i] = EventState::Free;
         events_interface.registered[i] = false;
     }
     auto nvmap_dev = std::make_shared<Devices::nvmap>(system);
     devices["/dev/nvhost-as-gpu"] = std::make_shared<Devices::nvhost_as_gpu>(system, nvmap_dev);
-    devices["/dev/nvhost-gpu"] = std::make_shared<Devices::nvhost_gpu>(system, nvmap_dev);
+    devices["/dev/nvhost-gpu"] =
+        std::make_shared<Devices::nvhost_gpu>(system, nvmap_dev, syncpoint_manager);
     devices["/dev/nvhost-ctrl-gpu"] = std::make_shared<Devices::nvhost_ctrl_gpu>(system);
     devices["/dev/nvmap"] = nvmap_dev;
     devices["/dev/nvdisp_disp0"] = std::make_shared<Devices::nvdisp_disp0>(system, nvmap_dev);
-    devices["/dev/nvhost-ctrl"] = std::make_shared<Devices::nvhost_ctrl>(system, events_interface);
-    devices["/dev/nvhost-nvdec"] = std::make_shared<Devices::nvhost_nvdec>(system);
+    devices["/dev/nvhost-ctrl"] =
+        std::make_shared<Devices::nvhost_ctrl>(system, events_interface, syncpoint_manager);
+    devices["/dev/nvhost-nvdec"] = std::make_shared<Devices::nvhost_nvdec>(system, nvmap_dev);
     devices["/dev/nvhost-nvjpg"] = std::make_shared<Devices::nvhost_nvjpg>(system);
-    devices["/dev/nvhost-vic"] = std::make_shared<Devices::nvhost_vic>(system);
+    devices["/dev/nvhost-vic"] = std::make_shared<Devices::nvhost_vic>(system, nvmap_dev);
 }
 
 Module::~Module() = default;
@@ -95,17 +98,17 @@ void Module::SignalSyncpt(const u32 syncpoint_id, const u32 value) {
         if (events_interface.assigned_syncpt[i] == syncpoint_id &&
             events_interface.assigned_value[i] == value) {
             events_interface.LiberateEvent(i);
-            events_interface.events[i].writable->Signal();
+            events_interface.events[i].event.writable->Signal();
         }
     }
 }
 
 std::shared_ptr<Kernel::ReadableEvent> Module::GetEvent(const u32 event_id) const {
-    return events_interface.events[event_id].readable;
+    return events_interface.events[event_id].event.readable;
 }
 
 std::shared_ptr<Kernel::WritableEvent> Module::GetEventWriteable(const u32 event_id) const {
-    return events_interface.events[event_id].writable;
+    return events_interface.events[event_id].event.writable;
 }
 
 } // namespace Service::Nvidia

src/core/hle/service/nvdrv/nvdrv.h

@@ -10,6 +10,7 @@
 #include "common/common_types.h"
 #include "core/hle/kernel/writable_event.h"
 #include "core/hle/service/nvdrv/nvdata.h"
+#include "core/hle/service/nvdrv/syncpoint_manager.h"
 #include "core/hle/service/service.h"
 
 namespace Core {
@@ -22,15 +23,23 @@ class NVFlinger;
 
 namespace Service::Nvidia {
 
+class SyncpointManager;
+
 namespace Devices {
 class nvdevice;
 }
 
+/// Represents an Nvidia event
+struct NvEvent {
+    Kernel::EventPair event;
+    Fence fence{};
+};
+
 struct EventInterface {
     // Mask representing currently busy events
     u64 events_mask{};
     // Each kernel event associated to an NV event
-    std::array<Kernel::EventPair, MaxNvEvents> events;
+    std::array<NvEvent, MaxNvEvents> events;
     // The status of the current NVEvent
     std::array<EventState, MaxNvEvents> status{};
     // Tells if an NVEvent is registered or not
@@ -119,6 +128,9 @@ public:
     std::shared_ptr<Kernel::WritableEvent> GetEventWriteable(u32 event_id) const;
 
 private:
+    /// Manages syncpoints on the host
+    SyncpointManager syncpoint_manager;
+
     /// Id to use for the next open file descriptor.
     u32 next_fd = 1;
 

src/core/hle/service/nvdrv/syncpoint_manager.cpp

@@ -0,0 +1,39 @@
+// Copyright 2020 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/assert.h"
+#include "core/hle/service/nvdrv/syncpoint_manager.h"
+#include "video_core/gpu.h"
+
+namespace Service::Nvidia {
+
+SyncpointManager::SyncpointManager(Tegra::GPU& gpu) : gpu{gpu} {}
+
+SyncpointManager::~SyncpointManager() = default;
+
+u32 SyncpointManager::RefreshSyncpoint(u32 syncpoint_id) {
+    syncpoints[syncpoint_id].min = gpu.GetSyncpointValue(syncpoint_id);
+    return GetSyncpointMin(syncpoint_id);
+}
+
+u32 SyncpointManager::AllocateSyncpoint() {
+    for (u32 syncpoint_id = 1; syncpoint_id < MaxSyncPoints; syncpoint_id++) {
+        if (!syncpoints[syncpoint_id].is_allocated) {
+            syncpoints[syncpoint_id].is_allocated = true;
+            return syncpoint_id;
+        }
+    }
+    UNREACHABLE_MSG("No more available syncpoints!");
+    return {};
+}
+
+u32 SyncpointManager::IncreaseSyncpoint(u32 syncpoint_id, u32 value) {
+    for (u32 index = 0; index < value; ++index) {
+        syncpoints[syncpoint_id].max.fetch_add(1, std::memory_order_relaxed);
+    }
+
+    return GetSyncpointMax(syncpoint_id);
+}
+
+} // namespace Service::Nvidia

src/core/hle/service/nvdrv/syncpoint_manager.h

@@ -0,0 +1,85 @@
+// Copyright 2020 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <atomic>
+
+#include "common/common_types.h"
+#include "core/hle/service/nvdrv/nvdata.h"
+
+namespace Tegra {
+class GPU;
+}
+
+namespace Service::Nvidia {
+
+class SyncpointManager final {
+public:
+    explicit SyncpointManager(Tegra::GPU& gpu);
+    ~SyncpointManager();
+
+    /**
+     * Returns true if the specified syncpoint is expired for the given value.
+     * @param syncpoint_id Syncpoint ID to check.
+     * @param value Value to check against the specified syncpoint.
+     * @returns True if the specified syncpoint is expired for the given value, otherwise False.
+     */
+    bool IsSyncpointExpired(u32 syncpoint_id, u32 value) const {
+        return (GetSyncpointMax(syncpoint_id) - value) >= (GetSyncpointMin(syncpoint_id) - value);
+    }
+
+    /**
+     * Gets the lower bound for the specified syncpoint.
+     * @param syncpoint_id Syncpoint ID to get the lower bound for.
+     * @returns The lower bound for the specified syncpoint.
+     */
+    u32 GetSyncpointMin(u32 syncpoint_id) const {
+        return syncpoints[syncpoint_id].min.load(std::memory_order_relaxed);
+    }
+
+    /**
+     * Gets the uper bound for the specified syncpoint.
+     * @param syncpoint_id Syncpoint ID to get the upper bound for.
+     * @returns The upper bound for the specified syncpoint.
+     */
+    u32 GetSyncpointMax(u32 syncpoint_id) const {
+        return syncpoints[syncpoint_id].max.load(std::memory_order_relaxed);
+    }
+
+    /**
+     * Refreshes the minimum value for the specified syncpoint.
+     * @param syncpoint_id Syncpoint ID to be refreshed.
+     * @returns The new syncpoint minimum value.
+     */
+    u32 RefreshSyncpoint(u32 syncpoint_id);
+
+    /**
+     * Allocates a new syncoint.
+     * @returns The syncpoint ID for the newly allocated syncpoint.
+     */
+    u32 AllocateSyncpoint();
+
+    /**
+     * Increases the maximum value for the specified syncpoint.
+     * @param syncpoint_id Syncpoint ID to be increased.
+     * @param value Value to increase the specified syncpoint by.
+     * @returns The new syncpoint maximum value.
+     */
+    u32 IncreaseSyncpoint(u32 syncpoint_id, u32 value);
+
+private:
+    struct Syncpoint {
+        std::atomic<u32> min;
+        std::atomic<u32> max;
+        std::atomic<bool> is_allocated;
+    };
+
+    std::array<Syncpoint, MaxSyncPoints> syncpoints{};
+
+    Tegra::GPU& gpu;
+};
+
+} // namespace Service::Nvidia

src/core/hle/service/nvflinger/buffer_queue.cpp

@@ -99,6 +99,20 @@ void BufferQueue::QueueBuffer(u32 slot, BufferTransformFlags transform,
     queue_sequence.push_back(slot);
 }
 
+void BufferQueue::CancelBuffer(u32 slot, const Service::Nvidia::MultiFence& multi_fence) {
+    const auto itr = std::find_if(queue.begin(), queue.end(),
+                                  [slot](const Buffer& buffer) { return buffer.slot == slot; });
+    ASSERT(itr != queue.end());
+    ASSERT(itr->status != Buffer::Status::Free);
+    itr->status = Buffer::Status::Free;
+    itr->multi_fence = multi_fence;
+    itr->swap_interval = 0;
+
+    free_buffers.push_back(slot);
+
+    buffer_wait_event.writable->Signal();
+}
+
 std::optional<std::reference_wrapper<const BufferQueue::Buffer>> BufferQueue::AcquireBuffer() {
     auto itr = queue.end();
     // Iterate to find a queued buffer matching the requested slot.

src/core/hle/service/nvflinger/buffer_queue.h

@@ -95,6 +95,7 @@ public:
     void QueueBuffer(u32 slot, BufferTransformFlags transform,
                      const Common::Rectangle<int>& crop_rect, u32 swap_interval,
                      Service::Nvidia::MultiFence& multi_fence);
+    void CancelBuffer(u32 slot, const Service::Nvidia::MultiFence& multi_fence);
     std::optional<std::reference_wrapper<const Buffer>> AcquireBuffer();
     void ReleaseBuffer(u32 slot);
     void Disconnect();

src/core/hle/service/nvflinger/nvflinger.cpp

@@ -242,6 +242,10 @@ void NVFlinger::Compose() {
 
         const auto& igbp_buffer = buffer->get().igbp_buffer;
 
+        if (!system.IsPoweredOn()) {
+            return; // We are likely shutting down
+        }
+
         auto& gpu = system.GPU();
         const auto& multi_fence = buffer->get().multi_fence;
         guard->unlock();

src/core/hle/service/set/set.cpp

@@ -202,6 +202,7 @@ SET::SET() : ServiceFramework("set") {
         {8, &SET::GetQuestFlag, "GetQuestFlag"},
         {9, &SET::GetKeyCodeMap2, "GetKeyCodeMap2"},
         {10, nullptr, "GetFirmwareVersionForDebug"},
+        {11, nullptr, "GetDeviceNickName"},
     };
     // clang-format on