51 files changed, 1198 insertions, 257 deletions

diff --git a/.travis-build.sh b/.travis-build.sh
index 22a3a9fd6..e06a4299b 100755
--- a/.travis-build.sh
+++ b/.travis-build.sh
@@ -11,8 +11,8 @@ fi
 
 #if OS is linux or is not set
 if [ "$TRAVIS_OS_NAME" = "linux" -o -z "$TRAVIS_OS_NAME" ]; then
-    export CC=gcc-4.9
-    export CXX=g++-4.9
+    export CC=gcc-5
+    export CXX=g++-5
     export PKG_CONFIG_PATH=$HOME/.local/lib/pkgconfig:$PKG_CONFIG_PATH
 
     mkdir build && cd build
diff --git a/.travis-deps.sh b/.travis-deps.sh
index eb99ead4f..c7bb7e785 100755
--- a/.travis-deps.sh
+++ b/.travis-deps.sh
@@ -5,8 +5,8 @@ set -x
 
 #if OS is linux or is not set
 if [ "$TRAVIS_OS_NAME" = "linux" -o -z "$TRAVIS_OS_NAME" ]; then
-    export CC=gcc-4.9
-    export CXX=g++-4.9
+    export CC=gcc-5
+    export CXX=g++-5
     mkdir -p $HOME/.local
 
     curl -L http://www.cmake.org/files/v2.8/cmake-2.8.11-Linux-i386.tar.gz \
diff --git a/.travis.yml b/.travis.yml
index 2e875cccf..8d86baece 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -15,8 +15,8 @@ addons:
     sources:
       - ubuntu-toolchain-r-test
     packages:
-      - gcc-4.9
-      - g++-4.9
+      - gcc-5
+      - g++-5
       - xorg-dev
       - lib32stdc++6 # For CMake
       - lftp # To upload builds
diff --git a/CMakeLists.txt b/CMakeLists.txt
index d6a4a915a..3a0a161e7 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -65,7 +65,7 @@ endif()
 message(STATUS "Target architecture: ${ARCHITECTURE}")
 
 if (NOT MSVC)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 -Wno-attributes -pthread")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++1y -Wno-attributes -pthread")
     set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -pthread")
 
     if (ARCHITECTURE_x86_64)
diff --git a/externals/cmake-modules/FindSDL2.cmake b/externals/cmake-modules/FindSDL2.cmake
index 0af86840a..9b8daa0d1 100644
--- a/externals/cmake-modules/FindSDL2.cmake
+++ b/externals/cmake-modules/FindSDL2.cmake
@@ -134,11 +134,17 @@ SET(SDL2_SEARCH_PATHS
     ${SDL2_PATH}
 )
 
+if(CMAKE_SIZEOF_VOID_P EQUAL 8)
+    set(VC_LIB_PATH_SUFFIX lib/x64)
+else()
+    set(VC_LIB_PATH_SUFFIX lib/x86)
+endif()
+
 FIND_LIBRARY(SDL2_LIBRARY_TEMP
     NAMES SDL2
     HINTS
     $ENV{SDL2DIR}
-    PATH_SUFFIXES lib64 lib
+    PATH_SUFFIXES lib64 lib ${VC_LIB_PATH_SUFFIX}
     PATHS ${SDL2_SEARCH_PATHS}
 )
 
diff --git a/src/audio_core/CMakeLists.txt b/src/audio_core/CMakeLists.txt
index b0d1c7eb6..869da5e83 100644
--- a/src/audio_core/CMakeLists.txt
+++ b/src/audio_core/CMakeLists.txt
@@ -1,12 +1,17 @@
 set(SRCS
             audio_core.cpp
+            codec.cpp
             hle/dsp.cpp
+            hle/filter.cpp
             hle/pipe.cpp
             )
 
 set(HEADERS
             audio_core.h
+            codec.h
+            hle/common.h
             hle/dsp.h
+            hle/filter.h
             hle/pipe.h
             sink.h
             )
diff --git a/src/audio_core/codec.cpp b/src/audio_core/codec.cpp
new file mode 100644
index 000000000..ab65514b7
--- /dev/null
+++ b/src/audio_core/codec.cpp
@@ -0,0 +1,122 @@
+// Copyright 2016 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <array>
+#include <cstddef>
+#include <cstring>
+#include <vector>
+
+#include "audio_core/codec.h"
+
+#include "common/assert.h"
+#include "common/common_types.h"
+#include "common/math_util.h"
+
+namespace Codec {
+
+StereoBuffer16 DecodeADPCM(const u8* const data, const size_t sample_count, const std::array<s16, 16>& adpcm_coeff, ADPCMState& state) {
+    // GC-ADPCM with scale factor and variable coefficients.
+    // Frames are 8 bytes long containing 14 samples each.
+    // Samples are 4 bits (one nibble) long.
+
+    constexpr size_t FRAME_LEN = 8;
+    constexpr size_t SAMPLES_PER_FRAME = 14;
+    constexpr std::array<int, 16> SIGNED_NIBBLES {{ 0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1 }};
+
+    const size_t ret_size = sample_count % 2 == 0 ? sample_count : sample_count + 1; // Ensure multiple of two.
+    StereoBuffer16 ret(ret_size);
+
+    int yn1 = state.yn1,
+        yn2 = state.yn2;
+
+    const size_t NUM_FRAMES = (sample_count + (SAMPLES_PER_FRAME - 1)) / SAMPLES_PER_FRAME; // Round up.
+    for (size_t framei = 0; framei < NUM_FRAMES; framei++) {
+        const int frame_header = data[framei * FRAME_LEN];
+        const int scale = 1 << (frame_header & 0xF);
+        const int idx = (frame_header >> 4) & 0x7;
+
+        // Coefficients are fixed point with 11 bits fractional part.
+        const int coef1 = adpcm_coeff[idx * 2 + 0];
+        const int coef2 = adpcm_coeff[idx * 2 + 1];
+
+        // Decodes an audio sample. One nibble produces one sample.
+        const auto decode_sample = [&](const int nibble) -> s16 {
+            const int xn = nibble * scale;
+            // We first transform everything into 11 bit fixed point, perform the second order digital filter, then transform back.
+            // 0x400 == 0.5 in 11 bit fixed point.
+            // Filter: y[n] = x[n] + 0.5 + c1 * y[n-1] + c2 * y[n-2]
+            int val = ((xn << 11) + 0x400 + coef1 * yn1 + coef2 * yn2) >> 11;
+            // Clamp to output range.
+            val = MathUtil::Clamp(val, -32768, 32767);
+            // Advance output feedback.
+            yn2 = yn1;
+            yn1 = val;
+            return (s16)val;
+        };
+
+        size_t outputi = framei * SAMPLES_PER_FRAME;
+        size_t datai = framei * FRAME_LEN + 1;
+        for (size_t i = 0; i < SAMPLES_PER_FRAME && outputi < sample_count; i += 2) {
+            const s16 sample1 = decode_sample(SIGNED_NIBBLES[data[datai] & 0xF]);
+            ret[outputi].fill(sample1);
+            outputi++;
+
+            const s16 sample2 = decode_sample(SIGNED_NIBBLES[data[datai] >> 4]);
+            ret[outputi].fill(sample2);
+            outputi++;
+
+            datai++;
+        }
+    }
+
+    state.yn1 = yn1;
+    state.yn2 = yn2;
+
+    return ret;
+}
+
+static s16 SignExtendS8(u8 x) {
+    // The data is actually signed PCM8.
+    // We sign extend this to signed PCM16.
+    return static_cast<s16>(static_cast<s8>(x));
+}
+
+StereoBuffer16 DecodePCM8(const unsigned num_channels, const u8* const data, const size_t sample_count) {
+    ASSERT(num_channels == 1 || num_channels == 2);
+
+    StereoBuffer16 ret(sample_count);
+
+    if (num_channels == 1) {
+        for (size_t i = 0; i < sample_count; i++) {
+            ret[i].fill(SignExtendS8(data[i]));
+        }
+    } else {
+        for (size_t i = 0; i < sample_count; i++) {
+            ret[i][0] = SignExtendS8(data[i * 2 + 0]);
+            ret[i][1] = SignExtendS8(data[i * 2 + 1]);
+        }
+    }
+
+    return ret;
+}
+
+StereoBuffer16 DecodePCM16(const unsigned num_channels, const u8* const data, const size_t sample_count) {
+    ASSERT(num_channels == 1 || num_channels == 2);
+
+    StereoBuffer16 ret(sample_count);
+
+    if (num_channels == 1) {
+        for (size_t i = 0; i < sample_count; i++) {
+            s16 sample;
+            std::memcpy(&sample, data + i * sizeof(s16), sizeof(s16));
+            ret[i].fill(sample);
+        }
+    } else {
+        std::memcpy(ret.data(), data, sample_count * 2 * sizeof(u16));
+    }
+
+    return ret;
+}
+
+};
diff --git a/src/audio_core/codec.h b/src/audio_core/codec.h
new file mode 100644
index 000000000..e695f2edc
--- /dev/null
+++ b/src/audio_core/codec.h
@@ -0,0 +1,50 @@
+// Copyright 2016 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <vector>
+
+#include "common/common_types.h"
+
+namespace Codec {
+
+/// A variable length buffer of signed PCM16 stereo samples.
+using StereoBuffer16 = std::vector<std::array<s16, 2>>;
+
+/// See: Codec::DecodeADPCM
+struct ADPCMState {
+    // Two historical samples from previous processed buffer,
+    // required for ADPCM decoding
+    s16 yn1; ///< y[n-1]
+    s16 yn2; ///< y[n-2]
+};
+
+/**
+ * @param data Pointer to buffer that contains ADPCM data to decode
+ * @param sample_count Length of buffer in terms of number of samples
+ * @param adpcm_coeff ADPCM coefficients
+ * @param state ADPCM state, this is updated with new state
+ * @return Decoded stereo signed PCM16 data, sample_count in length
+ */
+StereoBuffer16 DecodeADPCM(const u8* const data, const size_t sample_count, const std::array<s16, 16>& adpcm_coeff, ADPCMState& state);
+
+/**
+ * @param num_channels Number of channels
+ * @param data Pointer to buffer that contains PCM8 data to decode
+ * @param sample_count Length of buffer in terms of number of samples
+ * @return Decoded stereo signed PCM16 data, sample_count in length
+ */
+StereoBuffer16 DecodePCM8(const unsigned num_channels, const u8* const data, const size_t sample_count);
+
+/**
+ * @param num_channels Number of channels
+ * @param data Pointer to buffer that contains PCM16 data to decode
+ * @param sample_count Length of buffer in terms of number of samples
+ * @return Decoded stereo signed PCM16 data, sample_count in length
+ */
+StereoBuffer16 DecodePCM16(const unsigned num_channels, const u8* const data, const size_t sample_count);
+
+};
diff --git a/src/audio_core/hle/common.h b/src/audio_core/hle/common.h
new file mode 100644
index 000000000..37d441eb2
--- /dev/null
+++ b/src/audio_core/hle/common.h
@@ -0,0 +1,35 @@
+// Copyright 2016 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <algorithm>
+#include <array>
+
+#include "audio_core/audio_core.h"
+
+#include "common/common_types.h"
+
+namespace DSP {
+namespace HLE {
+
+/// The final output to the speakers is stereo. Preprocessing output in Source is also stereo.
+using StereoFrame16 = std::array<std::array<s16, 2>, AudioCore::samples_per_frame>;
+
+/// The DSP is quadraphonic internally.
+using QuadFrame32   = std::array<std::array<s32, 4>, AudioCore::samples_per_frame>;
+
+/**
+ * This performs the filter operation defined by FilterT::ProcessSample on the frame in-place.
+ * FilterT::ProcessSample is called sequentially on the samples.
+ */
+template<typename FrameT, typename FilterT>
+void FilterFrame(FrameT& frame, FilterT& filter) {
+    std::transform(frame.begin(), frame.end(), frame.begin(), [&filter](const typename FrameT::value_type& sample) {
+        return filter.ProcessSample(sample);
+    });
+}
+
+} // namespace HLE
+} // namespace DSP
diff --git a/src/audio_core/hle/dsp.h b/src/audio_core/hle/dsp.h
index 376436c29..c15ef0b7a 100644
--- a/src/audio_core/hle/dsp.h
+++ b/src/audio_core/hle/dsp.h
@@ -126,8 +126,11 @@ struct SourceConfiguration {
         union {
             u32_le dirty_raw;
 
+            BitField<0, 1, u32_le> format_dirty;
+            BitField<1, 1, u32_le> mono_or_stereo_dirty;
             BitField<2, 1, u32_le> adpcm_coefficients_dirty;
             BitField<3, 1, u32_le> partial_embedded_buffer_dirty; ///< Tends to be set when a looped buffer is queued.
+            BitField<4, 1, u32_le> partial_reset_flag;
 
             BitField<16, 1, u32_le> enable_dirty;
             BitField<17, 1, u32_le> interpolation_dirty;
@@ -143,8 +146,7 @@ struct SourceConfiguration {
             BitField<27, 1, u32_le> gain_2_dirty;
             BitField<28, 1, u32_le> sync_dirty;
             BitField<29, 1, u32_le> reset_flag;
-
-            BitField<31, 1, u32_le> embedded_buffer_dirty;
+            BitField<30, 1, u32_le> embedded_buffer_dirty;
         };
 
         // Gain control
@@ -175,7 +177,8 @@ struct SourceConfiguration {
         /**
          * This is the simplest normalized first-order digital recursive filter.
          * The transfer function of this filter is:
-         *     H(z) = b0 / (1 + a1 z^-1)
+         *     H(z) = b0 / (1 - a1 z^-1)
+         * Note the feedbackward coefficient is negated.
          * Values are signed fixed point with 15 fractional bits.
          */
         struct SimpleFilter {
@@ -192,11 +195,11 @@ struct SourceConfiguration {
          * Values are signed fixed point with 14 fractional bits.
          */
         struct BiquadFilter {
-            s16_le b0;
-            s16_le b1;
-            s16_le b2;
-            s16_le a1;
             s16_le a2;
+            s16_le a1;
+            s16_le b2;
+            s16_le b1;
+            s16_le b0;
         };
 
         union {
diff --git a/src/audio_core/hle/filter.cpp b/src/audio_core/hle/filter.cpp
new file mode 100644
index 000000000..2c65ef026
--- /dev/null
+++ b/src/audio_core/hle/filter.cpp
@@ -0,0 +1,115 @@
+// Copyright 2016 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <array>
+#include <cstddef>
+
+#include "audio_core/hle/common.h"
+#include "audio_core/hle/dsp.h"
+#include "audio_core/hle/filter.h"
+
+#include "common/common_types.h"
+#include "common/math_util.h"
+
+namespace DSP {
+namespace HLE {
+
+void SourceFilters::Reset() {
+    Enable(false, false);
+}
+
+void SourceFilters::Enable(bool simple, bool biquad) {
+    simple_filter_enabled = simple;
+    biquad_filter_enabled = biquad;
+
+    if (!simple)
+        simple_filter.Reset();
+    if (!biquad)
+        biquad_filter.Reset();
+}
+
+void SourceFilters::Configure(SourceConfiguration::Configuration::SimpleFilter config) {
+    simple_filter.Configure(config);
+}
+
+void SourceFilters::Configure(SourceConfiguration::Configuration::BiquadFilter config) {
+    biquad_filter.Configure(config);
+}
+
+void SourceFilters::ProcessFrame(StereoFrame16& frame) {
+    if (!simple_filter_enabled && !biquad_filter_enabled)
+        return;
+
+    if (simple_filter_enabled) {
+        FilterFrame(frame, simple_filter);
+    }
+
+    if (biquad_filter_enabled) {
+        FilterFrame(frame, biquad_filter);
+    }
+}
+
+// SimpleFilter
+
+void SourceFilters::SimpleFilter::Reset() {
+    y1.fill(0);
+    // Configure as passthrough.
+    a1 = 0;
+    b0 = 1 << 15;
+}
+
+void SourceFilters::SimpleFilter::Configure(SourceConfiguration::Configuration::SimpleFilter config) {
+    a1 = config.a1;
+    b0 = config.b0;
+}
+
+std::array<s16, 2> SourceFilters::SimpleFilter::ProcessSample(const std::array<s16, 2>& x0) {
+    std::array<s16, 2> y0;
+    for (size_t i = 0; i < 2; i++) {
+        const s32 tmp = (b0 * x0[i] + a1 * y1[i]) >> 15;
+        y0[i] = MathUtil::Clamp(tmp, -32768, 32767);
+    }
+
+    y1 = y0;
+
+    return y0;
+}
+
+// BiquadFilter
+
+void SourceFilters::BiquadFilter::Reset() {
+    x1.fill(0);
+    x2.fill(0);
+    y1.fill(0);
+    y2.fill(0);
+    // Configure as passthrough.
+    a1 = a2 = b1 = b2 = 0;
+    b0 = 1 << 14;
+}
+
+void SourceFilters::BiquadFilter::Configure(SourceConfiguration::Configuration::BiquadFilter config) {
+    a1 = config.a1;
+    a2 = config.a2;
+    b0 = config.b0;
+    b1 = config.b1;
+    b2 = config.b2;
+}
+
+std::array<s16, 2> SourceFilters::BiquadFilter::ProcessSample(const std::array<s16, 2>& x0) {
+    std::array<s16, 2> y0;
+    for (size_t i = 0; i < 2; i++) {
+        const s32 tmp = (b0 * x0[i] + b1 * x1[i] + b2 * x2[i] + a1 * y1[i] + a2 * y2[i]) >> 14;
+        y0[i] = MathUtil::Clamp(tmp, -32768, 32767);
+    }
+
+    x2 = x1;
+    x1 = x0;
+    y2 = y1;
+    y1 = y0;
+
+    return y0;
+}
+
+} // namespace HLE
+} // namespace DSP
diff --git a/src/audio_core/hle/filter.h b/src/audio_core/hle/filter.h
new file mode 100644
index 000000000..75738f600
--- /dev/null
+++ b/src/audio_core/hle/filter.h
@@ -0,0 +1,112 @@
+// Copyright 2016 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+
+#include "audio_core/hle/common.h"
+#include "audio_core/hle/dsp.h"
+
+#include "common/common_types.h"
+
+namespace DSP {
+namespace HLE {
+
+/// Preprocessing filters. There is an independent set of filters for each Source.
+class SourceFilters final {
+    SourceFilters() { Reset(); }
+
+    /// Reset internal state.
+    void Reset();
+
+    /**
+     * Enable/Disable filters
+     * See also: SourceConfiguration::Configuration::simple_filter_enabled,
+     *           SourceConfiguration::Configuration::biquad_filter_enabled.
+     * @param simple If true, enables the simple filter. If false, disables it.
+     * @param simple If true, enables the biquad filter. If false, disables it.
+     */
+    void Enable(bool simple, bool biquad);
+
+    /**
+     * Configure simple filter.
+     * @param config Configuration from DSP shared memory.
+     */
+    void Configure(SourceConfiguration::Configuration::SimpleFilter config);
+
+    /**
+     * Configure biquad filter.
+     * @param config Configuration from DSP shared memory.
+     */
+    void Configure(SourceConfiguration::Configuration::BiquadFilter config);
+
+    /**
+     * Processes a frame in-place.
+     * @param frame Audio samples to process. Modified in-place.
+     */
+    void ProcessFrame(StereoFrame16& frame);
+
+private:
+    bool simple_filter_enabled;
+    bool biquad_filter_enabled;
+
+    struct SimpleFilter {
+        SimpleFilter() { Reset(); }
+
+        /// Resets internal state.
+        void Reset();
+
+        /**
+         * Configures this filter with application settings.
+         * @param config Configuration from DSP shared memory.
+         */
+        void Configure(SourceConfiguration::Configuration::SimpleFilter config);
+
+        /**
+         * Processes a single stereo PCM16 sample.
+         * @param x0 Input sample
+         * @return Output sample
+         */
+        std::array<s16, 2> ProcessSample(const std::array<s16, 2>& x0);
+
+    private:
+        // Configuration
+        s32 a1, b0;
+        // Internal state
+        std::array<s16, 2> y1;
+    } simple_filter;
+
+    struct BiquadFilter {
+        BiquadFilter() { Reset(); }
+
+        /// Resets internal state.
+        void Reset();
+
+        /**
+         * Configures this filter with application settings.
+         * @param config Configuration from DSP shared memory.
+         */
+        void Configure(SourceConfiguration::Configuration::BiquadFilter config);
+
+        /**
+         * Processes a single stereo PCM16 sample.
+         * @param x0 Input sample
+         * @return Output sample
+         */
+        std::array<s16, 2> ProcessSample(const std::array<s16, 2>& x0);
+
+    private:
+        // Configuration
+        s32 a1, a2, b0, b1, b2;
+        // Internal state
+        std::array<s16, 2> x1;
+        std::array<s16, 2> x2;
+        std::array<s16, 2> y1;
+        std::array<s16, 2> y2;
+    } biquad_filter;
+};
+
+} // namespace HLE
+} // namespace DSP
diff --git a/src/citra_qt/game_list.cpp b/src/citra_qt/game_list.cpp
index 1f8d69a03..ffcab1f03 100644
--- a/src/citra_qt/game_list.cpp
+++ b/src/citra_qt/game_list.cpp
@@ -66,7 +66,7 @@ void GameList::ValidateEntry(const QModelIndex& item)
 
     if (file_path.isEmpty())
         return;
-    std::string std_file_path = file_path.toStdString();
+    std::string std_file_path(file_path.toStdString());
     if (!FileUtil::Exists(std_file_path) || FileUtil::IsDirectory(std_file_path))
         return;
     emit GameChosen(file_path);
diff --git a/src/citra_qt/main.cpp b/src/citra_qt/main.cpp
index ebd52a7b3..ca0ae6f7b 100644
--- a/src/citra_qt/main.cpp
+++ b/src/citra_qt/main.cpp
@@ -417,7 +417,7 @@ void GMainWindow::UpdateRecentFiles() {
 }
 
 void GMainWindow::OnGameListLoadFile(QString game_path) {
-    BootGame(game_path.toLocal8Bit().data());
+    BootGame(game_path.toStdString());
 }
 
 void GMainWindow::OnMenuLoadFile() {
@@ -428,7 +428,7 @@ void GMainWindow::OnMenuLoadFile() {
     if (!filename.isEmpty()) {
         settings.setValue("romsPath", QFileInfo(filename).path());
 
-        BootGame(filename.toLocal8Bit().data());
+        BootGame(filename.toStdString());
     }
 }
 
@@ -440,7 +440,7 @@ void GMainWindow::OnMenuLoadSymbolMap() {
     if (!filename.isEmpty()) {
         settings.setValue("symbolsPath", QFileInfo(filename).path());
 
-        LoadSymbolMap(filename.toLocal8Bit().data());
+        LoadSymbolMap(filename.toStdString());
     }
 }
 
@@ -461,7 +461,7 @@ void GMainWindow::OnMenuRecentFile() {
     QString filename = action->data().toString();
     QFileInfo file_info(filename);
     if (file_info.exists()) {
-        BootGame(filename.toLocal8Bit().data());
+        BootGame(filename.toStdString());
     } else {
         // Display an error message and remove the file from the list.
         QMessageBox::information(this, tr("File not found"), tr("File \"%1\" not found").arg(filename));
diff --git a/src/common/emu_window.h b/src/common/emu_window.h
index a0ae4c9fa..7c3486dea 100644
--- a/src/common/emu_window.h
+++ b/src/common/emu_window.h
@@ -105,7 +105,7 @@ public:
      * @todo Fix this function to be thread-safe.
      * @return PadState object indicating the current pad state
      */
-    const Service::HID::PadState GetPadState() const {
+    Service::HID::PadState GetPadState() const {
         return pad_state;
     }
 
@@ -116,11 +116,59 @@ public:
      * @return std::tuple of (x, y, pressed) where `x` and `y` are the touch coordinates and
      *         `pressed` is true if the touch screen is currently being pressed
      */
-    const std::tuple<u16, u16, bool> GetTouchState() const {
+    std::tuple<u16, u16, bool> GetTouchState() const {
         return std::make_tuple(touch_x, touch_y, touch_pressed);
     }
 
     /**
+     * Gets the current accelerometer state (acceleration along each three axis).
+     * Axis explained:
+     *   +x is the same direction as LEFT on D-pad.
+     *   +y is normal to the touch screen, pointing outward.
+     *   +z is the same direction as UP on D-pad.
+     * Units:
+     *   1 unit of return value = 1/512 g (measured by hw test),
+     *   where g is the gravitational acceleration (9.8 m/sec2).
+     * @note This should be called by the core emu thread to get a state set by the window thread.
+     * @todo Implement accelerometer input in front-end.
+     * @return std::tuple of (x, y, z)
+     */
+    std::tuple<s16, s16, s16> GetAccelerometerState() const {
+        // stubbed
+        return std::make_tuple(0, -512, 0);
+    }
+
+    /**
+     * Gets the current gyroscope state (angular rates about each three axis).
+     * Axis explained:
+     *   +x is the same direction as LEFT on D-pad.
+     *   +y is normal to the touch screen, pointing outward.
+     *   +z is the same direction as UP on D-pad.
+     * Orientation is determined by right-hand rule.
+     * Units:
+     *   1 unit of return value = (1/coef) deg/sec,
+     *   where coef is the return value of GetGyroscopeRawToDpsCoefficient().
+     * @note This should be called by the core emu thread to get a state set by the window thread.
+     * @todo Implement gyroscope input in front-end.
+     * @return std::tuple of (x, y, z)
+     */
+    std::tuple<s16, s16, s16> GetGyroscopeState() const {
+        // stubbed
+        return std::make_tuple(0, 0, 0);
+    }
+
+    /**
+     * Gets the coefficient for units conversion of gyroscope state.
+     * The conversion formula is r = coefficient * v,
+     * where v is angular rate in deg/sec,
+     * and r is the gyroscope state.
+     * @return float-type coefficient
+     */
+    f32 GetGyroscopeRawToDpsCoefficient() const {
+        return 14.375f; // taken from hw test, and gyroscope's document
+    }
+
+    /**
      * Returns currently active configuration.
      * @note Accesses to the returned object need not be consistent because it may be modified in another thread
      */
diff --git a/src/common/file_util.cpp b/src/common/file_util.cpp
index c3061479a..9ada09f8a 100644
--- a/src/common/file_util.cpp
+++ b/src/common/file_util.cpp
@@ -85,7 +85,7 @@ bool Exists(const std::string &filename)
     StripTailDirSlashes(copy);
 
 #ifdef _WIN32
-    int result = _tstat64(Common::UTF8ToTStr(copy).c_str(), &file_info);
+    int result = _wstat64(Common::UTF8ToUTF16W(copy).c_str(), &file_info);
 #else
     int result = stat64(copy.c_str(), &file_info);
 #endif
@@ -102,7 +102,7 @@ bool IsDirectory(const std::string &filename)
     StripTailDirSlashes(copy);
 
 #ifdef _WIN32
-    int result = _tstat64(Common::UTF8ToTStr(copy).c_str(), &file_info);
+    int result = _wstat64(Common::UTF8ToUTF16W(copy).c_str(), &file_info);
 #else
     int result = stat64(copy.c_str(), &file_info);
 #endif
@@ -138,7 +138,7 @@ bool Delete(const std::string &filename)
     }
 
 #ifdef _WIN32
-    if (!DeleteFile(Common::UTF8ToTStr(filename).c_str()))
+    if (!DeleteFileW(Common::UTF8ToUTF16W(filename).c_str()))
     {
         LOG_ERROR(Common_Filesystem, "DeleteFile failed on %s: %s",
                  filename.c_str(), GetLastErrorMsg());
@@ -160,7 +160,7 @@ bool CreateDir(const std::string &path)
 {
     LOG_TRACE(Common_Filesystem, "directory %s", path.c_str());
 #ifdef _WIN32
-    if (::CreateDirectory(Common::UTF8ToTStr(path).c_str(), nullptr))
+    if (::CreateDirectoryW(Common::UTF8ToUTF16W(path).c_str(), nullptr))
         return true;
     DWORD error = GetLastError();
     if (error == ERROR_ALREADY_EXISTS)
@@ -241,7 +241,7 @@ bool DeleteDir(const std::string &filename)
     }
 
 #ifdef _WIN32
-    if (::RemoveDirectory(Common::UTF8ToTStr(filename).c_str()))
+    if (::RemoveDirectoryW(Common::UTF8ToUTF16W(filename).c_str()))
         return true;
 #else
     if (rmdir(filename.c_str()) == 0)
@@ -257,8 +257,13 @@ bool Rename(const std::string &srcFilename, const std::string &destFilename)
 {
     LOG_TRACE(Common_Filesystem, "%s --> %s",
             srcFilename.c_str(), destFilename.c_str());
+#ifdef _WIN32
+    if (_wrename(Common::UTF8ToUTF16W(srcFilename).c_str(), Common::UTF8ToUTF16W(destFilename).c_str()) == 0)
+        return true;
+#else
     if (rename(srcFilename.c_str(), destFilename.c_str()) == 0)
         return true;
+#endif
     LOG_ERROR(Common_Filesystem, "failed %s --> %s: %s",
               srcFilename.c_str(), destFilename.c_str(), GetLastErrorMsg());
     return false;
@@ -270,7 +275,7 @@ bool Copy(const std::string &srcFilename, const std::string &destFilename)
     LOG_TRACE(Common_Filesystem, "%s --> %s",
             srcFilename.c_str(), destFilename.c_str());
 #ifdef _WIN32
-    if (CopyFile(Common::UTF8ToTStr(srcFilename).c_str(), Common::UTF8ToTStr(destFilename).c_str(), FALSE))
+    if (CopyFileW(Common::UTF8ToUTF16W(srcFilename).c_str(), Common::UTF8ToUTF16W(destFilename).c_str(), FALSE))
         return true;
 
     LOG_ERROR(Common_Filesystem, "failed %s --> %s: %s",
@@ -358,7 +363,7 @@ u64 GetSize(const std::string &filename)
 
     struct stat64 buf;
 #ifdef _WIN32
-    if (_tstat64(Common::UTF8ToTStr(filename).c_str(), &buf) == 0)
+    if (_wstat64(Common::UTF8ToUTF16W(filename).c_str(), &buf) == 0)
 #else
     if (stat64(filename.c_str(), &buf) == 0)
 #endif
@@ -432,16 +437,16 @@ bool ForeachDirectoryEntry(unsigned* num_entries_out, const std::string &directo
 
 #ifdef _WIN32
     // Find the first file in the directory.
-    WIN32_FIND_DATA ffd;
+    WIN32_FIND_DATAW ffd;
 
-    HANDLE handle_find = FindFirstFile(Common::UTF8ToTStr(directory + "\\*").c_str(), &ffd);
+    HANDLE handle_find = FindFirstFileW(Common::UTF8ToUTF16W(directory + "\\*").c_str(), &ffd);
     if (handle_find == INVALID_HANDLE_VALUE) {
         FindClose(handle_find);
         return false;
     }
     // windows loop
     do {
-        const std::string virtual_name(Common::TStrToUTF8(ffd.cFileName));
+        const std::string virtual_name(Common::UTF16ToUTF8(ffd.cFileName));
 #else
     struct dirent dirent, *result = nullptr;
 
@@ -465,7 +470,7 @@ bool ForeachDirectoryEntry(unsigned* num_entries_out, const std::string &directo
         found_entries += ret_entries;
 
 #ifdef _WIN32
-    } while (FindNextFile(handle_find, &ffd) != 0);
+    } while (FindNextFileW(handle_find, &ffd) != 0);
     FindClose(handle_find);
 #else
     }
@@ -572,15 +577,23 @@ void CopyDir(const std::string &source_path, const std::string &dest_path)
 // Returns the current directory
 std::string GetCurrentDir()
 {
-    char *dir;
     // Get the current working directory (getcwd uses malloc)
+#ifdef _WIN32
+    wchar_t *dir;
+    if (!(dir = _wgetcwd(nullptr, 0))) {
+#else
+    char *dir;
     if (!(dir = getcwd(nullptr, 0))) {
-
+#endif
         LOG_ERROR(Common_Filesystem, "GetCurrentDirectory failed: %s",
                 GetLastErrorMsg());
         return nullptr;
     }
+#ifdef _WIN32
+    std::string strDir = Common::UTF16ToUTF8(dir);
+#else
     std::string strDir = dir;
+#endif
     free(dir);
     return strDir;
 }
@@ -588,7 +601,11 @@ std::string GetCurrentDir()
 // Sets the current directory to the given directory
 bool SetCurrentDir(const std::string &directory)
 {
+#ifdef _WIN32
+    return _wchdir(Common::UTF8ToUTF16W(directory).c_str()) == 0;
+#else
     return chdir(directory.c_str()) == 0;
+#endif
 }
 
 #if defined(__APPLE__)
@@ -613,9 +630,9 @@ std::string& GetExeDirectory()
     static std::string exe_path;
     if (exe_path.empty())
     {
-        TCHAR tchar_exe_path[2048];
-        GetModuleFileName(nullptr, tchar_exe_path, 2048);
-        exe_path = Common::TStrToUTF8(tchar_exe_path);
+        wchar_t wchar_exe_path[2048];
+        GetModuleFileNameW(nullptr, wchar_exe_path, 2048);
+        exe_path = Common::UTF16ToUTF8(wchar_exe_path);
         exe_path = exe_path.substr(0, exe_path.find_last_of('\\'));
     }
     return exe_path;
@@ -900,7 +917,7 @@ bool IOFile::Open(const std::string& filename, const char openmode[])
 {
     Close();
 #ifdef _WIN32
-    _tfopen_s(&m_file, Common::UTF8ToTStr(filename).c_str(), Common::UTF8ToTStr(openmode).c_str());
+    _wfopen_s(&m_file, Common::UTF8ToUTF16W(filename).c_str(), Common::UTF8ToUTF16W(openmode).c_str());
 #else
     m_file = fopen(filename.c_str(), openmode);
 #endif
diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp
index 4c86151ab..cfbfbc2a7 100644
--- a/src/common/logging/backend.cpp
+++ b/src/common/logging/backend.cpp
@@ -46,8 +46,10 @@ namespace Log {
         SUB(Service, NIM) \
         SUB(Service, NWM) \
         SUB(Service, CAM) \
+        SUB(Service, CECD) \
         SUB(Service, CFG) \
         SUB(Service, DSP) \
+        SUB(Service, DLP) \
         SUB(Service, HID) \
         SUB(Service, SOC) \
         SUB(Service, IR) \
diff --git a/src/common/logging/log.h b/src/common/logging/log.h
index e4c39c308..4f6856f3d 100644
--- a/src/common/logging/log.h
+++ b/src/common/logging/log.h
@@ -61,8 +61,10 @@ enum class Class : ClassType {
     Service_NIM,                ///< The NIM (Network interface manager) service
     Service_NWM,                ///< The NWM (Network wlan manager) service
     Service_CAM,                ///< The CAM (Camera) service
+    Service_CECD,               ///< The CECD service
     Service_CFG,                ///< The CFG (Configuration) service
     Service_DSP,                ///< The DSP (DSP control) service
+    Service_DLP,                ///< The DLP (Download Play) service
     Service_HID,                ///< The HID (Human interface device) service
     Service_SOC,                ///< The SOC (Socket) service
     Service_IR,                 ///< The IR service
diff --git a/src/common/string_util.cpp b/src/common/string_util.cpp
index 6d6fc591f..f0aa072db 100644
--- a/src/common/string_util.cpp
+++ b/src/common/string_util.cpp
@@ -320,27 +320,27 @@ std::u16string UTF8ToUTF16(const std::string& input)
 #endif
 }
 
-static std::string UTF16ToUTF8(const std::wstring& input)
+static std::wstring CPToUTF16(u32 code_page, const std::string& input)
 {
-    auto const size = WideCharToMultiByte(CP_UTF8, 0, input.data(), static_cast<int>(input.size()), nullptr, 0, nullptr, nullptr);
+    auto const size = MultiByteToWideChar(code_page, 0, input.data(), static_cast<int>(input.size()), nullptr, 0);
 
-    std::string output;
+    std::wstring output;
     output.resize(size);
 
-    if (size == 0 || size != WideCharToMultiByte(CP_UTF8, 0, input.data(), static_cast<int>(input.size()), &output[0], static_cast<int>(output.size()), nullptr, nullptr))
+    if (size == 0 || size != MultiByteToWideChar(code_page, 0, input.data(), static_cast<int>(input.size()), &output[0], static_cast<int>(output.size())))
         output.clear();
 
     return output;
 }
 
-static std::wstring CPToUTF16(u32 code_page, const std::string& input)
+std::string UTF16ToUTF8(const std::wstring& input)
 {
-    auto const size = MultiByteToWideChar(code_page, 0, input.data(), static_cast<int>(input.size()), nullptr, 0);
+    auto const size = WideCharToMultiByte(CP_UTF8, 0, input.data(), static_cast<int>(input.size()), nullptr, 0, nullptr, nullptr);
 
-    std::wstring output;
+    std::string output;
     output.resize(size);
 
-    if (size == 0 || size != MultiByteToWideChar(code_page, 0, input.data(), static_cast<int>(input.size()), &output[0], static_cast<int>(output.size())))
+    if (size == 0 || size != WideCharToMultiByte(CP_UTF8, 0, input.data(), static_cast<int>(input.size()), &output[0], static_cast<int>(output.size()), nullptr, nullptr))
         output.clear();
 
     return output;
diff --git a/src/common/string_util.h b/src/common/string_util.h
index c5c474c6f..89d9f133e 100644
--- a/src/common/string_util.h
+++ b/src/common/string_util.h
@@ -95,7 +95,7 @@ std::string CP1252ToUTF8(const std::string& str);
 std::string SHIFTJISToUTF8(const std::string& str);
 
 #ifdef _WIN32
-
+std::string UTF16ToUTF8(const std::wstring& input);
 std::wstring UTF8ToUTF16W(const std::string& str);
 
 #ifdef _UNICODE
diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt
index 3473e2f5b..a8d891689 100644
--- a/src/core/CMakeLists.txt
+++ b/src/core/CMakeLists.txt
@@ -68,6 +68,7 @@ set(SRCS
             hle/service/cfg/cfg_s.cpp
             hle/service/cfg/cfg_u.cpp
             hle/service/csnd_snd.cpp
+            hle/service/dlp_srvr.cpp
             hle/service/dsp_dsp.cpp
             hle/service/err_f.cpp
             hle/service/frd/frd.cpp
@@ -200,6 +201,7 @@ set(HEADERS
             hle/service/cfg/cfg_s.h
             hle/service/cfg/cfg_u.h
             hle/service/csnd_snd.h
+            hle/service/dlp_srvr.h
             hle/service/dsp_dsp.h
             hle/service/err_f.h
             hle/service/frd/frd.h
diff --git a/src/core/arm/dyncom/arm_dyncom_interpreter.cpp b/src/core/arm/dyncom/arm_dyncom_interpreter.cpp
index 5f8826034..9ed61947e 100644
--- a/src/core/arm/dyncom/arm_dyncom_interpreter.cpp
+++ b/src/core/arm/dyncom/arm_dyncom_interpreter.cpp
@@ -36,7 +36,8 @@ enum {
     CALL            = (1 << 4),
     RET             = (1 << 5),
     END_OF_PAGE     = (1 << 6),
-    THUMB           = (1 << 7)
+    THUMB           = (1 << 7),
+    SINGLE_STEP     = (1 << 8)
 };
 
 #define RM    BITS(sht_oper, 0, 3)
@@ -3466,7 +3467,35 @@ enum {
 
 MICROPROFILE_DEFINE(DynCom_Decode, "DynCom", "Decode", MP_RGB(255, 64, 64));
 
-static int InterpreterTranslate(ARMul_State* cpu, int& bb_start, u32 addr) {
+static unsigned int InterpreterTranslateInstruction(const ARMul_State* cpu, const u32 phys_addr, ARM_INST_PTR& inst_base) {
+    unsigned int inst_size = 4;
+    unsigned int inst = Memory::Read32(phys_addr & 0xFFFFFFFC);
+
+    // If we are in Thumb mode, we'll translate one Thumb instruction to the corresponding ARM instruction
+    if (cpu->TFlag) {
+        u32 arm_inst;
+        ThumbDecodeStatus state = DecodeThumbInstruction(inst, phys_addr, &arm_inst, &inst_size, &inst_base);
+
+        // We have translated the Thumb branch instruction in the Thumb decoder
+        if (state == ThumbDecodeStatus::BRANCH) {
+            return inst_size;
+        }
+        inst = arm_inst;
+    }
+
+    int idx;
+    if (DecodeARMInstruction(inst, &idx) == ARMDecodeStatus::FAILURE) {
+        std::string disasm = ARM_Disasm::Disassemble(phys_addr, inst);
+        LOG_ERROR(Core_ARM11, "Decode failure.\tPC : [0x%x]\tInstruction : %s [%x]", phys_addr, disasm.c_str(), inst);
+        LOG_ERROR(Core_ARM11, "cpsr=0x%x, cpu->TFlag=%d, r15=0x%x", cpu->Cpsr, cpu->TFlag, cpu->Reg[15]);
+        CITRA_IGNORE_EXIT(-1);
+    }
+    inst_base = arm_instruction_trans[idx](inst, idx);
+
+    return inst_size;
+}
+
+static int InterpreterTranslateBlock(ARMul_State* cpu, int& bb_start, u32 addr) {
     Common::Profiling::ScopeTimer timer_decode(profile_decode);
     MICROPROFILE_SCOPE(DynCom_Decode);
 
@@ -3475,8 +3504,6 @@ static int InterpreterTranslate(ARMul_State* cpu, int& bb_start, u32 addr) {
     // Go on next, until terminal instruction
     // Save start addr of basicblock in CreamCache
     ARM_INST_PTR inst_base = nullptr;
-    unsigned int inst, inst_size = 4;
-    int idx;
     int ret = NON_BRANCH;
     int size = 0; // instruction size of basic block
     bb_start = top;
@@ -3485,30 +3512,10 @@ static int InterpreterTranslate(ARMul_State* cpu, int& bb_start, u32 addr) {
     u32 pc_start = cpu->Reg[15];
 
     while (ret == NON_BRANCH) {
-        inst = Memory::Read32(phys_addr & 0xFFFFFFFC);
+        unsigned int inst_size = InterpreterTranslateInstruction(cpu, phys_addr, inst_base);
 
         size++;
-        // If we are in Thumb mode, we'll translate one Thumb instruction to the corresponding ARM instruction
-        if (cpu->TFlag) {
-            u32 arm_inst;
-            ThumbDecodeStatus state = DecodeThumbInstruction(inst, phys_addr, &arm_inst, &inst_size, &inst_base);
-
-            // We have translated the Thumb branch instruction in the Thumb decoder
-            if (state == ThumbDecodeStatus::BRANCH) {
-                goto translated;
-            }
-            inst = arm_inst;
-        }
-
-        if (DecodeARMInstruction(inst, &idx) == ARMDecodeStatus::FAILURE) {
-            std::string disasm = ARM_Disasm::Disassemble(phys_addr, inst);
-            LOG_ERROR(Core_ARM11, "Decode failure.\tPC : [0x%x]\tInstruction : %s [%x]", phys_addr, disasm.c_str(), inst);
-            LOG_ERROR(Core_ARM11, "cpsr=0x%x, cpu->TFlag=%d, r15=0x%x", cpu->Cpsr, cpu->TFlag, cpu->Reg[15]);
-            CITRA_IGNORE_EXIT(-1);
-        }
-        inst_base = arm_instruction_trans[idx](inst, idx);
 
-translated:
         phys_addr += inst_size;
 
         if ((phys_addr & 0xfff) == 0) {
@@ -3522,6 +3529,27 @@ translated:
     return KEEP_GOING;
 }
 
+static int InterpreterTranslateSingle(ARMul_State* cpu, int& bb_start, u32 addr) {
+    Common::Profiling::ScopeTimer timer_decode(profile_decode);
+    MICROPROFILE_SCOPE(DynCom_Decode);
+
+    ARM_INST_PTR inst_base = nullptr;
+    bb_start = top;
+
+    u32 phys_addr = addr;
+    u32 pc_start = cpu->Reg[15];
+
+    InterpreterTranslateInstruction(cpu, phys_addr, inst_base);
+
+    if (inst_base->br == NON_BRANCH) {
+        inst_base->br = SINGLE_STEP;
+    }
+
+    cpu->instruction_cache[pc_start] = bb_start;
+
+    return KEEP_GOING;
+}
+
 static int clz(unsigned int x) {
     int n;
     if (x == 0) return (32);
@@ -3871,8 +3899,11 @@ unsigned InterpreterMainLoop(ARMul_State* cpu) {
         auto itr = cpu->instruction_cache.find(cpu->Reg[15]);
         if (itr != cpu->instruction_cache.end()) {
             ptr = itr->second;
+        } else if (cpu->NumInstrsToExecute != 1) {
+            if (InterpreterTranslateBlock(cpu, ptr, cpu->Reg[15]) == FETCH_EXCEPTION)
+                goto END;
         } else {
-            if (InterpreterTranslate(cpu, ptr, cpu->Reg[15]) == FETCH_EXCEPTION)
+            if (InterpreterTranslateSingle(cpu, ptr, cpu->Reg[15]) == FETCH_EXCEPTION)
                 goto END;
         }
 
diff --git a/src/core/arm/skyeye_common/armstate.cpp b/src/core/arm/skyeye_common/armstate.cpp
index 2d814345a..5550c112e 100644
--- a/src/core/arm/skyeye_common/armstate.cpp
+++ b/src/core/arm/skyeye_common/armstate.cpp
@@ -2,6 +2,7 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <algorithm>
 #include "common/swap.h"
 #include "common/logging/log.h"
 #include "core/memory.h"
@@ -48,8 +49,7 @@ void ARMul_State::ChangePrivilegeMode(u32 new_mode)
             Spsr[UNDEFBANK] = Spsr_copy;
             break;
         case FIQ32MODE:
-            Reg_firq[0] = Reg[13];
-            Reg_firq[1] = Reg[14];
+            std::copy(Reg.begin() + 8, Reg.end() - 1, Reg_firq.begin());
             Spsr[FIQBANK] = Spsr_copy;
             break;
         }
@@ -85,8 +85,7 @@ void ARMul_State::ChangePrivilegeMode(u32 new_mode)
             Bank = UNDEFBANK;
             break;
         case FIQ32MODE:
-            Reg[13] = Reg_firq[0];
-            Reg[14] = Reg_firq[1];
+            std::copy(Reg_firq.begin(), Reg_firq.end(), Reg.begin() + 8);
             Spsr_copy = Spsr[FIQBANK];
             Bank = FIQBANK;
             break;
diff --git a/src/core/file_sys/archive_extsavedata.h b/src/core/file_sys/archive_extsavedata.h
index 287a6fee1..e9a72850d 100644
--- a/src/core/file_sys/archive_extsavedata.h
+++ b/src/core/file_sys/archive_extsavedata.h
@@ -45,13 +45,14 @@ public:
     void WriteIcon(const Path& path, const u8* icon_data, size_t icon_size);
 
 private:
+    bool shared; ///< Whether this archive represents an ExtSaveData archive or a SharedExtSaveData archive
+
     /**
      * This holds the full directory path for this archive, it is only set after a successful call
      * to Open, this is formed as <base extsavedatapath>/<type>/<high>/<low>.
      * See GetExtSaveDataPath for the code that extracts this data from an archive path.
      */
     std::string mount_point;
-    bool shared; ///< Whether this archive represents an ExtSaveData archive or a SharedExtSaveData archive
 };
 
 /**
diff --git a/src/core/hle/kernel/session.h b/src/core/hle/kernel/session.h
index adaffcafe..6ddaf970e 100644
--- a/src/core/hle/kernel/session.h
+++ b/src/core/hle/kernel/session.h
@@ -16,23 +16,23 @@
 
 namespace IPC {
 
-inline u32 MakeHeader(u16 command_id, unsigned int regular_params, unsigned int translate_params) {
+constexpr u32 MakeHeader(u16 command_id, unsigned int regular_params, unsigned int translate_params) {
     return ((u32)command_id << 16) | (((u32)regular_params & 0x3F) << 6) | (((u32)translate_params & 0x3F) << 0);
 }
 
-inline u32 MoveHandleDesc(unsigned int num_handles = 1) {
+constexpr u32 MoveHandleDesc(unsigned int num_handles = 1) {
     return 0x0 | ((num_handles - 1) << 26);
 }
 
-inline u32 CopyHandleDesc(unsigned int num_handles = 1) {
+constexpr u32 CopyHandleDesc(unsigned int num_handles = 1) {
     return 0x10 | ((num_handles - 1) << 26);
 }
 
-inline u32 CallingPidDesc() {
+constexpr u32 CallingPidDesc() {
     return 0x20;
 }
 
-inline u32 StaticBufferDesc(u32 size, unsigned int buffer_id) {
+constexpr u32 StaticBufferDesc(u32 size, unsigned int buffer_id) {
     return 0x2 | (size << 14) | ((buffer_id & 0xF) << 10);
 }
 
@@ -42,7 +42,7 @@ enum MappedBufferPermissions {
     RW = R | W,
 };
 
-inline u32 MappedBufferDesc(u32 size, MappedBufferPermissions perms) {
+constexpr u32 MappedBufferDesc(u32 size, MappedBufferPermissions perms) {
     return 0x8 | (size << 4) | (u32)perms;
 }
 
diff --git a/src/core/hle/result.h b/src/core/hle/result.h
index 0cb76ba1c..2d22652d9 100644
--- a/src/core/hle/result.h
+++ b/src/core/hle/result.h
@@ -24,6 +24,7 @@ enum class ErrorDescription : u32 {
     FS_InvalidOpenFlags = 230,
     FS_NotAFile = 250,
     FS_NotFormatted = 340, ///< This is used by the FS service when creating a SaveData archive
+    OutofRangeOrMisalignedAddress = 513, // TODO(purpasmart): Check if this name fits its actual usage
     FS_InvalidPath = 702,
     InvalidSection = 1000,
     TooLarge = 1001,
diff --git a/src/core/hle/service/cecd/cecd.cpp b/src/core/hle/service/cecd/cecd.cpp
index 6d79ce9b4..e6e36e7ec 100644
--- a/src/core/hle/service/cecd/cecd.cpp
+++ b/src/core/hle/service/cecd/cecd.cpp
@@ -4,6 +4,7 @@
 
 #include "common/logging/log.h"
 
+#include "core/hle/kernel/event.h"
 #include "core/hle/service/service.h"
 #include "core/hle/service/cecd/cecd.h"
 #include "core/hle/service/cecd/cecd_s.h"
@@ -12,14 +13,38 @@
 namespace Service {
 namespace CECD {
 
-void Init() {
-    using namespace Kernel;
+static Kernel::SharedPtr<Kernel::Event> cecinfo_event;
+static Kernel::SharedPtr<Kernel::Event> change_state_event;
+
+void GetCecInfoEventHandle(Service::Interface* self) {
+    u32* cmd_buff = Kernel::GetCommandBuffer();
+
+    cmd_buff[1] = RESULT_SUCCESS.raw; // No error
+    cmd_buff[3] = Kernel::g_handle_table.Create(cecinfo_event).MoveFrom(); // Event handle
+
+    LOG_WARNING(Service_CECD, "(STUBBED) called");
+}
+
+void GetChangeStateEventHandle(Service::Interface* self) {
+    u32* cmd_buff = Kernel::GetCommandBuffer();
 
+    cmd_buff[1] = RESULT_SUCCESS.raw; // No error
+    cmd_buff[3] = Kernel::g_handle_table.Create(change_state_event).MoveFrom(); // Event handle
+
+    LOG_WARNING(Service_CECD, "(STUBBED) called");
+}
+
+void Init() {
     AddService(new CECD_S_Interface);
     AddService(new CECD_U_Interface);
+
+    cecinfo_event = Kernel::Event::Create(Kernel::ResetType::OneShot, "CECD_U::cecinfo_event");
+    change_state_event = Kernel::Event::Create(Kernel::ResetType::OneShot, "CECD_U::change_state_event");
 }
 
 void Shutdown() {
+    cecinfo_event = nullptr;
+    change_state_event = nullptr;
 }
 
 } // namespace CECD
diff --git a/src/core/hle/service/cecd/cecd.h b/src/core/hle/service/cecd/cecd.h
index 9e158521b..89a8d67bb 100644
--- a/src/core/hle/service/cecd/cecd.h
+++ b/src/core/hle/service/cecd/cecd.h
@@ -5,8 +5,31 @@
 #pragma once
 
 namespace Service {
+
+class Interface;
+
 namespace CECD {
 
+/**
+ * GetCecInfoEventHandle service function
+ *  Inputs:
+ *      0: 0x000F0000
+ *  Outputs:
+ *      1: ResultCode
+ *      3: Event Handle
+ */
+void GetCecInfoEventHandle(Service::Interface* self);
+
+/**
+ * GetChangeStateEventHandle service function
+ *  Inputs:
+ *      0: 0x00100000
+ *  Outputs:
+ *      1: ResultCode
+ *      3: Event Handle
+ */
+void GetChangeStateEventHandle(Service::Interface* self);
+
 /// Initialize CECD service(s)
 void Init();
 
diff --git a/src/core/hle/service/cecd/cecd_u.cpp b/src/core/hle/service/cecd/cecd_u.cpp
index 9b720a738..ace1c73c0 100644
--- a/src/core/hle/service/cecd/cecd_u.cpp
+++ b/src/core/hle/service/cecd/cecd_u.cpp
@@ -2,13 +2,16 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include "core/hle/service/cecd/cecd.h"
 #include "core/hle/service/cecd/cecd_u.h"
 
 namespace Service {
 namespace CECD {
 
 static const Interface::FunctionInfo FunctionTable[] = {
-    { 0x00120104, nullptr, "ReadSavedData" },
+    {0x000F0000, GetCecInfoEventHandle,     "GetCecInfoEventHandle"},
+    {0x00100000, GetChangeStateEventHandle, "GetChangeStateEventHandle"},
+    {0x00120104, nullptr,                   "ReadSavedData"},
 };
 
 CECD_U_Interface::CECD_U_Interface() {
diff --git a/src/core/hle/service/cfg/cfg_i.cpp b/src/core/hle/service/cfg/cfg_i.cpp
index 0559a07b2..b18060f6d 100644
--- a/src/core/hle/service/cfg/cfg_i.cpp
+++ b/src/core/hle/service/cfg/cfg_i.cpp
@@ -9,6 +9,18 @@ namespace Service {
 namespace CFG {
 
 const Interface::FunctionInfo FunctionTable[] = {
+    // cfg common
+    {0x00010082, GetConfigInfoBlk2,                    "GetConfigInfoBlk2"},
+    {0x00020000, SecureInfoGetRegion,                  "SecureInfoGetRegion"},
+    {0x00030040, GenHashConsoleUnique,                 "GenHashConsoleUnique"},
+    {0x00040000, GetRegionCanadaUSA,                   "GetRegionCanadaUSA"},
+    {0x00050000, GetSystemModel,                       "GetSystemModel"},
+    {0x00060000, GetModelNintendo2DS,                  "GetModelNintendo2DS"},
+    {0x00070040, nullptr,                              "WriteToFirstByteCfgSavegame"},
+    {0x00080080, nullptr,                              "GoThroughTable"},
+    {0x00090040, GetCountryCodeString,                 "GetCountryCodeString"},
+    {0x000A0040, GetCountryCodeID,                     "GetCountryCodeID"},
+    // cfg:i
     {0x04010082, GetConfigInfoBlk8,                    "GetConfigInfoBlk8"},
     {0x04020082, nullptr,                              "SetConfigInfoBlk4"},
     {0x04030000, UpdateConfigNANDSavegame,             "UpdateConfigNANDSavegame"},
diff --git a/src/core/hle/service/cfg/cfg_s.cpp b/src/core/hle/service/cfg/cfg_s.cpp
index b03d290e5..e001f7687 100644
--- a/src/core/hle/service/cfg/cfg_s.cpp
+++ b/src/core/hle/service/cfg/cfg_s.cpp
@@ -9,10 +9,18 @@ namespace Service {
 namespace CFG {
 
 const Interface::FunctionInfo FunctionTable[] = {
+    // cfg common
     {0x00010082, GetConfigInfoBlk2,                    "GetConfigInfoBlk2"},
     {0x00020000, SecureInfoGetRegion,                  "SecureInfoGetRegion"},
     {0x00030040, GenHashConsoleUnique,                 "GenHashConsoleUnique"},
+    {0x00040000, GetRegionCanadaUSA,                   "GetRegionCanadaUSA"},
     {0x00050000, GetSystemModel,                       "GetSystemModel"},
+    {0x00060000, GetModelNintendo2DS,                  "GetModelNintendo2DS"},
+    {0x00070040, nullptr,                              "WriteToFirstByteCfgSavegame"},
+    {0x00080080, nullptr,                              "GoThroughTable"},
+    {0x00090040, GetCountryCodeString,                 "GetCountryCodeString"},
+    {0x000A0040, GetCountryCodeID,                     "GetCountryCodeID"},
+    // cfg:s
     {0x04010082, GetConfigInfoBlk8,                    "GetConfigInfoBlk8"},
     {0x04020082, nullptr,                              "SetConfigInfoBlk4"},
     {0x04030000, UpdateConfigNANDSavegame,             "UpdateConfigNANDSavegame"},
diff --git a/src/core/hle/service/cfg/cfg_u.cpp b/src/core/hle/service/cfg/cfg_u.cpp
index 89ae96c9e..606f7b2eb 100644
--- a/src/core/hle/service/cfg/cfg_u.cpp
+++ b/src/core/hle/service/cfg/cfg_u.cpp
@@ -9,6 +9,7 @@ namespace Service {
 namespace CFG {
 
 const Interface::FunctionInfo FunctionTable[] = {
+    // cfg common
     {0x00010082, GetConfigInfoBlk2,     "GetConfigInfoBlk2"},
     {0x00020000, SecureInfoGetRegion,   "SecureInfoGetRegion"},
     {0x00030040, GenHashConsoleUnique,  "GenHashConsoleUnique"},
diff --git a/src/core/hle/service/dlp_srvr.cpp b/src/core/hle/service/dlp_srvr.cpp
new file mode 100644
index 000000000..1f30188da
--- /dev/null
+++ b/src/core/hle/service/dlp_srvr.cpp
@@ -0,0 +1,36 @@
+// Copyright 2016 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/logging/log.h"
+#include "core/hle/hle.h"
+#include "core/hle/service/dlp_srvr.h"
+
+////////////////////////////////////////////////////////////////////////////////////////////////////
+// Namespace DLP_SRVR
+
+namespace DLP_SRVR {
+
+static void unk_0x000E0040(Service::Interface* self) {
+    u32* cmd_buff = Kernel::GetCommandBuffer();
+
+    cmd_buff[1] = RESULT_SUCCESS.raw;
+    cmd_buff[2] = 0;
+
+    LOG_WARNING(Service_DLP, "(STUBBED) called");
+}
+
+const Interface::FunctionInfo FunctionTable[] = {
+    {0x00010183, nullptr,           "Initialize"},
+    {0x00020000, nullptr,           "Finalize"},
+    {0x000E0040, unk_0x000E0040,    "unk_0x000E0040"},
+};
+
+////////////////////////////////////////////////////////////////////////////////////////////////////
+// Interface class
+
+Interface::Interface() {
+    Register(FunctionTable);
+}
+
+} // namespace
diff --git a/src/core/hle/service/dlp_srvr.h b/src/core/hle/service/dlp_srvr.h
new file mode 100644
index 000000000..d65d00814
--- /dev/null
+++ b/src/core/hle/service/dlp_srvr.h
@@ -0,0 +1,23 @@
+// Copyright 2016 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/service/service.h"
+
+////////////////////////////////////////////////////////////////////////////////////////////////////
+// Namespace DLP_SRVR
+
+namespace DLP_SRVR {
+
+class Interface : public Service::Interface {
+public:
+    Interface();
+
+    std::string GetPortName() const override {
+        return "dlp:SRVR";
+    }
+};
+
+} // namespace
diff --git a/src/core/hle/service/gsp_gpu.cpp b/src/core/hle/service/gsp_gpu.cpp
index 2ace2cade..0c655395e 100644
--- a/src/core/hle/service/gsp_gpu.cpp
+++ b/src/core/hle/service/gsp_gpu.cpp
@@ -31,6 +31,13 @@ const static u32 REGS_BEGIN = 0x1EB00000;
 
 namespace GSP_GPU {
 
+const ResultCode ERR_GSP_REGS_OUTOFRANGE_OR_MISALIGNED(ErrorDescription::OutofRangeOrMisalignedAddress, ErrorModule::GX,
+    ErrorSummary::InvalidArgument, ErrorLevel::Usage); // 0xE0E02A01
+const ResultCode ERR_GSP_REGS_MISALIGNED(ErrorDescription::MisalignedSize, ErrorModule::GX,
+    ErrorSummary::InvalidArgument, ErrorLevel::Usage); // 0xE0E02BF2
+const ResultCode ERR_GSP_REGS_INVALID_SIZE(ErrorDescription::InvalidSize, ErrorModule::GX,
+    ErrorSummary::InvalidArgument, ErrorLevel::Usage); // 0xE0E02BEC
+
 /// Event triggered when GSP interrupt has been signalled
 Kernel::SharedPtr<Kernel::Event> g_interrupt_event;
 /// GSP shared memoryings
@@ -59,47 +66,87 @@ static inline InterruptRelayQueue* GetInterruptRelayQueue(u32 thread_id) {
 }
 
 /**
- * Checks if the parameters in a register write call are valid and logs in the case that
- * they are not
- * @param base_address The first address in the sequence of registers that will be written
- * @param size_in_bytes The number of registers that will be written
- * @return true if the parameters are valid, false otherwise
+ * Writes sequential GSP GPU hardware registers using an array of source data
+ *
+ * @param base_address The address of the first register in the sequence
+ * @param size_in_bytes The number of registers to update (size of data)
+ * @param data A pointer to the source data
+ * @return RESULT_SUCCESS if the parameters are valid, error code otherwise
  */
-static bool CheckWriteParameters(u32 base_address, u32 size_in_bytes) {
-    // TODO: Return proper error codes
-    if (base_address + size_in_bytes >= 0x420000) {
-        LOG_ERROR(Service_GSP, "Write address out of range! (address=0x%08x, size=0x%08x)",
+static ResultCode WriteHWRegs(u32 base_address, u32 size_in_bytes, const u32* data) {
+    // This magic number is verified to be done by the gsp module
+    const u32 max_size_in_bytes = 0x80;
+
+    if (base_address & 3 || base_address >= 0x420000) {
+        LOG_ERROR(Service_GSP, "Write address was out of range or misaligned! (address=0x%08x, size=0x%08x)",
                   base_address, size_in_bytes);
-        return false;
-    }
+        return ERR_GSP_REGS_OUTOFRANGE_OR_MISALIGNED;
+    } else if (size_in_bytes <= max_size_in_bytes) {
+        if (size_in_bytes & 3) {
+            LOG_ERROR(Service_GSP, "Misaligned size 0x%08x", size_in_bytes);
+            return ERR_GSP_REGS_MISALIGNED;
+        } else {
+            while (size_in_bytes > 0) {
+                HW::Write<u32>(base_address + REGS_BEGIN, *data);
+
+                size_in_bytes -= 4;
+                ++data;
+                base_address += 4;
+            }
+            return RESULT_SUCCESS;
+        }
 
-    // size should be word-aligned
-    if ((size_in_bytes % 4) != 0) {
-        LOG_ERROR(Service_GSP, "Invalid size 0x%08x", size_in_bytes);
-        return false;
+    } else {
+        LOG_ERROR(Service_GSP, "Out of range size 0x%08x", size_in_bytes);
+        return ERR_GSP_REGS_INVALID_SIZE;
     }
-
-    return true;
 }
 
 /**
- * Writes sequential GSP GPU hardware registers using an array of source data
+ * Updates sequential GSP GPU hardware registers using parallel arrays of source data and masks.
+ * For each register, the value is updated only where the mask is high
  *
  * @param base_address The address of the first register in the sequence
  * @param size_in_bytes The number of registers to update (size of data)
- * @param data A pointer to the source data
+ * @param data A pointer to the source data to use for updates
+ * @param masks A pointer to the masks
+ * @return RESULT_SUCCESS if the parameters are valid, error code otherwise
  */
-static void WriteHWRegs(u32 base_address, u32 size_in_bytes, const u32* data) {
-    // TODO: Return proper error codes
-    if (!CheckWriteParameters(base_address, size_in_bytes))
-        return;
+static ResultCode WriteHWRegsWithMask(u32 base_address, u32 size_in_bytes, const u32* data, const u32* masks) {
+    // This magic number is verified to be done by the gsp module
+    const u32 max_size_in_bytes = 0x80;
 
-    while (size_in_bytes > 0) {
-        HW::Write<u32>(base_address + REGS_BEGIN, *data);
+    if (base_address & 3 || base_address >= 0x420000) {
+        LOG_ERROR(Service_GSP, "Write address was out of range or misaligned! (address=0x%08x, size=0x%08x)",
+                  base_address, size_in_bytes);
+        return ERR_GSP_REGS_OUTOFRANGE_OR_MISALIGNED;
+    } else if (size_in_bytes <= max_size_in_bytes) {
+        if (size_in_bytes & 3) {
+            LOG_ERROR(Service_GSP, "Misaligned size 0x%08x", size_in_bytes);
+            return ERR_GSP_REGS_MISALIGNED;
+        } else {
+            while (size_in_bytes > 0) {
+                const u32 reg_address = base_address + REGS_BEGIN;
+
+                u32 reg_value;
+                HW::Read<u32>(reg_value, reg_address);
+
+                // Update the current value of the register only for set mask bits
+                reg_value = (reg_value & ~*masks) | (*data | *masks);
+
+                HW::Write<u32>(reg_address, reg_value);
+
+                size_in_bytes -= 4;
+                ++data;
+                ++masks;
+                base_address += 4;
+            }
+            return RESULT_SUCCESS;
+        }
 
-        size_in_bytes -= 4;
-        ++data;
-        base_address += 4;
+    } else {
+        LOG_ERROR(Service_GSP, "Out of range size 0x%08x", size_in_bytes);
+        return ERR_GSP_REGS_INVALID_SIZE;
     }
 }
 
@@ -120,39 +167,7 @@ static void WriteHWRegs(Service::Interface* self) {
 
     u32* src = (u32*)Memory::GetPointer(cmd_buff[4]);
 
-    WriteHWRegs(reg_addr, size, src);
-}
-
-/**
- * Updates sequential GSP GPU hardware registers using parallel arrays of source data and masks.
- * For each register, the value is updated only where the mask is high
- *
- * @param base_address The address of the first register in the sequence
- * @param size_in_bytes The number of registers to update (size of data)
- * @param data A pointer to the source data to use for updates
- * @param masks A pointer to the masks
- */
-static void WriteHWRegsWithMask(u32 base_address, u32 size_in_bytes, const u32* data, const u32* masks) {
-    // TODO: Return proper error codes
-    if (!CheckWriteParameters(base_address, size_in_bytes))
-        return;
-
-    while (size_in_bytes > 0) {
-        const u32 reg_address = base_address + REGS_BEGIN;
-
-        u32 reg_value;
-        HW::Read<u32>(reg_value, reg_address);
-
-        // Update the current value of the register only for set mask bits
-        reg_value = (reg_value & ~*masks) | (*data | *masks);
-
-        HW::Write<u32>(reg_address, reg_value);
-
-        size_in_bytes -= 4;
-        ++data;
-        ++masks;
-        base_address += 4;
-    }
+    cmd_buff[1] = WriteHWRegs(reg_addr, size, src).raw;
 }
 
 /**
@@ -174,7 +189,7 @@ static void WriteHWRegsWithMask(Service::Interface* self) {
     u32* src_data = (u32*)Memory::GetPointer(cmd_buff[4]);
     u32* mask_data = (u32*)Memory::GetPointer(cmd_buff[6]);
 
-    WriteHWRegsWithMask(reg_addr, size, src_data, mask_data);
+    cmd_buff[1] = WriteHWRegsWithMask(reg_addr, size, src_data, mask_data).raw;
 }
 
 /// Read a GSP GPU hardware register
@@ -206,27 +221,27 @@ static void ReadHWRegs(Service::Interface* self) {
     }
 }
 
-void SetBufferSwap(u32 screen_id, const FrameBufferInfo& info) {
+ResultCode SetBufferSwap(u32 screen_id, const FrameBufferInfo& info) {
     u32 base_address = 0x400000;
     PAddr phys_address_left = Memory::VirtualToPhysicalAddress(info.address_left);
     PAddr phys_address_right = Memory::VirtualToPhysicalAddress(info.address_right);
     if (info.active_fb == 0) {
-        WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].address_left1)), 4,
-                &phys_address_left);
-        WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].address_right1)), 4,
-                &phys_address_right);
+        WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].address_left1)),
+                    4, &phys_address_left);
+        WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].address_right1)),
+                    4, &phys_address_right);
     } else {
-        WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].address_left2)), 4,
-                &phys_address_left);
-        WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].address_right2)), 4,
-                &phys_address_right);
+        WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].address_left2)),
+                    4, &phys_address_left);
+        WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].address_right2)),
+                    4, &phys_address_right);
     }
-    WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].stride)), 4,
-            &info.stride);
-    WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].color_format)), 4,
-            &info.format);
-    WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].active_fb)), 4,
-            &info.shown_fb);
+    WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].stride)),
+                4, &info.stride);
+    WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].color_format)),
+                4, &info.format);
+    WriteHWRegs(base_address + 4 * static_cast<u32>(GPU_REG_INDEX(framebuffer_config[screen_id].active_fb)),
+                4, &info.shown_fb);
 
     if (Pica::g_debug_context)
         Pica::g_debug_context->OnEvent(Pica::DebugContext::Event::BufferSwapped, nullptr);
@@ -234,6 +249,8 @@ void SetBufferSwap(u32 screen_id, const FrameBufferInfo& info) {
     if (screen_id == 0) {
         MicroProfileFlip();
     }
+
+    return RESULT_SUCCESS;
 }
 
 /**
@@ -251,9 +268,8 @@ static void SetBufferSwap(Service::Interface* self) {
     u32* cmd_buff = Kernel::GetCommandBuffer();
     u32 screen_id = cmd_buff[1];
     FrameBufferInfo* fb_info = (FrameBufferInfo*)&cmd_buff[2];
-    SetBufferSwap(screen_id, *fb_info);
 
-    cmd_buff[1] = 0; // No error
+    cmd_buff[1] = SetBufferSwap(screen_id, *fb_info).raw;
 }
 
 /**
@@ -286,6 +302,22 @@ static void FlushDataCache(Service::Interface* self) {
 }
 
 /**
+ * GSP_GPU::SetAxiConfigQoSMode service function
+ *  Inputs:
+ *      1 : Mode, unused in emulator
+ *  Outputs:
+ *      1 : Result of function, 0 on success, otherwise error code
+ */
+static void SetAxiConfigQoSMode(Service::Interface* self) {
+    u32* cmd_buff = Kernel::GetCommandBuffer();
+    u32 mode = cmd_buff[1];
+
+    cmd_buff[1] = RESULT_SUCCESS.raw; // No error
+
+    LOG_WARNING(Service_GSP, "(STUBBED) called mode=0x%08X", mode);
+}
+
+/**
  * GSP_GPU::RegisterInterruptRelayQueue service function
  *  Inputs:
  *      1 : "Flags" field, purpose is unknown
@@ -302,6 +334,12 @@ static void RegisterInterruptRelayQueue(Service::Interface* self) {
     g_interrupt_event = Kernel::g_handle_table.Get<Kernel::Event>(cmd_buff[3]);
     ASSERT_MSG((g_interrupt_event != nullptr), "handle is not valid!");
 
+    g_interrupt_event->name = "GSP_GPU::interrupt_event";
+
+    using Kernel::MemoryPermission;
+    g_shared_memory = Kernel::SharedMemory::Create(0x1000, MemoryPermission::ReadWrite,
+        MemoryPermission::ReadWrite, "GSPSharedMem");
+
     Handle shmem_handle = Kernel::g_handle_table.Create(g_shared_memory).MoveFrom();
 
     // This specific code is required for a successful initialization, rather than 0
@@ -314,6 +352,22 @@ static void RegisterInterruptRelayQueue(Service::Interface* self) {
 }
 
 /**
+ * GSP_GPU::UnregisterInterruptRelayQueue service function
+ *  Outputs:
+ *      1 : Result of function, 0 on success, otherwise error code
+ */
+static void UnregisterInterruptRelayQueue(Service::Interface* self) {
+    u32* cmd_buff = Kernel::GetCommandBuffer();
+
+    g_shared_memory = nullptr;
+    g_interrupt_event = nullptr;
+
+    cmd_buff[1] = RESULT_SUCCESS.raw;
+
+    LOG_WARNING(Service_GSP, "called");
+}
+
+/**
  * Signals that the specified interrupt type has occurred to userland code
  * @param interrupt_id ID of interrupt that is being signalled
  * @todo This should probably take a thread_id parameter and only signal this thread?
@@ -591,11 +645,11 @@ const Interface::FunctionInfo FunctionTable[] = {
     {0x000D0140, nullptr,                       "SetDisplayTransfer"},
     {0x000E0180, nullptr,                       "SetTextureCopy"},
     {0x000F0200, nullptr,                       "SetMemoryFill"},
-    {0x00100040, nullptr,                       "SetAxiConfigQoSMode"},
+    {0x00100040, SetAxiConfigQoSMode,           "SetAxiConfigQoSMode"},
     {0x00110040, nullptr,                       "SetPerfLogMode"},
     {0x00120000, nullptr,                       "GetPerfLog"},
     {0x00130042, RegisterInterruptRelayQueue,   "RegisterInterruptRelayQueue"},
-    {0x00140000, nullptr,                       "UnregisterInterruptRelayQueue"},
+    {0x00140000, UnregisterInterruptRelayQueue, "UnregisterInterruptRelayQueue"},
     {0x00150002, nullptr,                       "TryAcquireRight"},
     {0x00160042, nullptr,                       "AcquireRight"},
     {0x00170000, nullptr,                       "ReleaseRight"},
@@ -616,10 +670,7 @@ Interface::Interface() {
     Register(FunctionTable);
 
     g_interrupt_event = nullptr;
-
-    using Kernel::MemoryPermission;
-    g_shared_memory = Kernel::SharedMemory::Create(0x1000, MemoryPermission::ReadWrite,
-            MemoryPermission::ReadWrite, "GSPSharedMem");
+    g_shared_memory = nullptr;
 
     g_thread_id = 0;
 }
diff --git a/src/core/hle/service/gsp_gpu.h b/src/core/hle/service/gsp_gpu.h
index 0e2f7a21e..55a993bb8 100644
--- a/src/core/hle/service/gsp_gpu.h
+++ b/src/core/hle/service/gsp_gpu.h
@@ -194,7 +194,7 @@ public:
  */
 void SignalInterrupt(InterruptId interrupt_id);
 
-void SetBufferSwap(u32 screen_id, const FrameBufferInfo& info);
+ResultCode SetBufferSwap(u32 screen_id, const FrameBufferInfo& info);
 
 /**
  * Retrieves the framebuffer info stored in the GSP shared memory for the
diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp
index cb4fd38e2..1053d0f40 100644
--- a/src/core/hle/service/hid/hid.cpp
+++ b/src/core/hle/service/hid/hid.cpp
@@ -33,6 +33,11 @@ static Kernel::SharedPtr<Kernel::Event> event_debug_pad;
 
 static u32 next_pad_index;
 static u32 next_touch_index;
+static u32 next_accelerometer_index;
+static u32 next_gyroscope_index;
+
+static int enable_accelerometer_count = 0; // positive means enabled
+static int enable_gyroscope_count = 0; // positive means enabled
 
 const std::array<Service::HID::PadState, Settings::NativeInput::NUM_INPUTS> pad_mapping = {{
     Service::HID::PAD_A, Service::HID::PAD_B, Service::HID::PAD_X, Service::HID::PAD_Y,
@@ -78,17 +83,17 @@ void Update() {
     PadState changed = { { (state.hex ^ old_state.hex) } };
 
     // Get the current Pad entry
-    PadDataEntry* pad_entry = &mem->pad.entries[mem->pad.index];
+    PadDataEntry& pad_entry = mem->pad.entries[mem->pad.index];
 
     // Update entry properties
-    pad_entry->current_state.hex = state.hex;
-    pad_entry->delta_additions.hex = changed.hex & state.hex;
-    pad_entry->delta_removals.hex = changed.hex & old_state.hex;;
+    pad_entry.current_state.hex = state.hex;
+    pad_entry.delta_additions.hex = changed.hex & state.hex;
+    pad_entry.delta_removals.hex = changed.hex & old_state.hex;;
 
     // Set circle Pad
-    pad_entry->circle_pad_x = state.circle_left  ? -MAX_CIRCLEPAD_POS :
+    pad_entry.circle_pad_x = state.circle_left  ? -MAX_CIRCLEPAD_POS :
                               state.circle_right ?  MAX_CIRCLEPAD_POS : 0x0;
-    pad_entry->circle_pad_y = state.circle_down  ? -MAX_CIRCLEPAD_POS :
+    pad_entry.circle_pad_y = state.circle_down  ? -MAX_CIRCLEPAD_POS :
                               state.circle_up    ?  MAX_CIRCLEPAD_POS : 0x0;
 
     // If we just updated index 0, provide a new timestamp
@@ -101,11 +106,11 @@ void Update() {
     next_touch_index = (next_touch_index + 1) % mem->touch.entries.size();
 
     // Get the current touch entry
-    TouchDataEntry* touch_entry = &mem->touch.entries[mem->touch.index];
+    TouchDataEntry& touch_entry = mem->touch.entries[mem->touch.index];
     bool pressed = false;
 
-    std::tie(touch_entry->x, touch_entry->y, pressed) = VideoCore::g_emu_window->GetTouchState();
-    touch_entry->valid.Assign(pressed ? 1 : 0);
+    std::tie(touch_entry.x, touch_entry.y, pressed) = VideoCore::g_emu_window->GetTouchState();
+    touch_entry.valid.Assign(pressed ? 1 : 0);
 
     // TODO(bunnei): We're not doing anything with offset 0xA8 + 0x18 of HID SharedMemory, which
     // supposedly is "Touch-screen entry, which contains the raw coordinate data prior to being
@@ -120,6 +125,58 @@ void Update() {
     // Signal both handles when there's an update to Pad or touch
     event_pad_or_touch_1->Signal();
     event_pad_or_touch_2->Signal();
+
+    // Update accelerometer
+    if (enable_accelerometer_count > 0) {
+        mem->accelerometer.index = next_accelerometer_index;
+        next_accelerometer_index = (next_accelerometer_index + 1) % mem->accelerometer.entries.size();
+
+        AccelerometerDataEntry& accelerometer_entry = mem->accelerometer.entries[mem->accelerometer.index];
+        std::tie(accelerometer_entry.x, accelerometer_entry.y, accelerometer_entry.z)
+            = VideoCore::g_emu_window->GetAccelerometerState();
+
+        // Make up "raw" entry
+        // TODO(wwylele):
+        // From hardware testing, the raw_entry values are approximately,
+        // but not exactly, as twice as corresponding entries (or with a minus sign).
+        // It may caused by system calibration to the accelerometer.
+        // Figure out how it works, or, if no game reads raw_entry,
+        // the following three lines can be removed and leave raw_entry unimplemented.
+        mem->accelerometer.raw_entry.x = -2 * accelerometer_entry.x;
+        mem->accelerometer.raw_entry.z = 2 * accelerometer_entry.y;
+        mem->accelerometer.raw_entry.y = -2 * accelerometer_entry.z;
+
+        // If we just updated index 0, provide a new timestamp
+        if (mem->accelerometer.index == 0) {
+            mem->accelerometer.index_reset_ticks_previous = mem->accelerometer.index_reset_ticks;
+            mem->accelerometer.index_reset_ticks = (s64)CoreTiming::GetTicks();
+        }
+
+        event_accelerometer->Signal();
+    }
+
+    // Update gyroscope
+    if (enable_gyroscope_count > 0) {
+        mem->gyroscope.index = next_gyroscope_index;
+        next_gyroscope_index = (next_gyroscope_index + 1) % mem->gyroscope.entries.size();
+
+        GyroscopeDataEntry& gyroscope_entry = mem->gyroscope.entries[mem->gyroscope.index];
+        std::tie(gyroscope_entry.x, gyroscope_entry.y, gyroscope_entry.z)
+            = VideoCore::g_emu_window->GetGyroscopeState();
+
+        // Make up "raw" entry
+        mem->gyroscope.raw_entry.x = gyroscope_entry.x;
+        mem->gyroscope.raw_entry.z = -gyroscope_entry.y;
+        mem->gyroscope.raw_entry.y = gyroscope_entry.z;
+
+        // If we just updated index 0, provide a new timestamp
+        if (mem->gyroscope.index == 0) {
+            mem->gyroscope.index_reset_ticks_previous = mem->gyroscope.index_reset_ticks;
+            mem->gyroscope.index_reset_ticks = (s64)CoreTiming::GetTicks();
+        }
+
+        event_gyroscope->Signal();
+    }
 }
 
 void GetIPCHandles(Service::Interface* self) {
@@ -139,40 +196,69 @@ void GetIPCHandles(Service::Interface* self) {
 void EnableAccelerometer(Service::Interface* self) {
     u32* cmd_buff = Kernel::GetCommandBuffer();
 
+    ++enable_accelerometer_count;
     event_accelerometer->Signal();
 
     cmd_buff[1] = RESULT_SUCCESS.raw;
 
-    LOG_WARNING(Service_HID, "(STUBBED) called");
+    LOG_DEBUG(Service_HID, "called");
 }
 
 void DisableAccelerometer(Service::Interface* self) {
     u32* cmd_buff = Kernel::GetCommandBuffer();
 
+    --enable_accelerometer_count;
     event_accelerometer->Signal();
 
     cmd_buff[1] = RESULT_SUCCESS.raw;
 
-    LOG_WARNING(Service_HID, "(STUBBED) called");
+    LOG_DEBUG(Service_HID, "called");
 }
 
 void EnableGyroscopeLow(Service::Interface* self) {
     u32* cmd_buff = Kernel::GetCommandBuffer();
 
+    ++enable_gyroscope_count;
     event_gyroscope->Signal();
 
     cmd_buff[1] = RESULT_SUCCESS.raw;
 
-    LOG_WARNING(Service_HID, "(STUBBED) called");
+    LOG_DEBUG(Service_HID, "called");
 }
 
 void DisableGyroscopeLow(Service::Interface* self) {
     u32* cmd_buff = Kernel::GetCommandBuffer();
 
+    --enable_gyroscope_count;
     event_gyroscope->Signal();
 
     cmd_buff[1] = RESULT_SUCCESS.raw;
 
+    LOG_DEBUG(Service_HID, "called");
+}
+
+void GetGyroscopeLowRawToDpsCoefficient(Service::Interface* self) {
+    u32* cmd_buff = Kernel::GetCommandBuffer();
+
+    cmd_buff[1] = RESULT_SUCCESS.raw;
+
+    f32 coef = VideoCore::g_emu_window->GetGyroscopeRawToDpsCoefficient();
+    memcpy(&cmd_buff[2], &coef, 4);
+}
+
+void GetGyroscopeLowCalibrateParam(Service::Interface* self) {
+    u32* cmd_buff = Kernel::GetCommandBuffer();
+
+    cmd_buff[1] = RESULT_SUCCESS.raw;
+
+    const s16 param_unit = 6700; // an approximate value taken from hw
+    GyroscopeCalibrateParam param = {
+        { 0, param_unit, -param_unit },
+        { 0, param_unit, -param_unit },
+        { 0, param_unit, -param_unit },
+    };
+    memcpy(&cmd_buff[2], &param, sizeof(param));
+
     LOG_WARNING(Service_HID, "(STUBBED) called");
 }
 
diff --git a/src/core/hle/service/hid/hid.h b/src/core/hle/service/hid/hid.h
index 517f4f2ae..170d19ea8 100644
--- a/src/core/hle/service/hid/hid.h
+++ b/src/core/hle/service/hid/hid.h
@@ -78,6 +78,24 @@ struct TouchDataEntry {
 };
 
 /**
+ * Structure of a single entry of accelerometer state history within HID shared memory
+ */
+struct AccelerometerDataEntry {
+    s16 x;
+    s16 y;
+    s16 z;
+};
+
+/**
+ * Structure of a single entry of gyroscope state history within HID shared memory
+ */
+struct GyroscopeDataEntry {
+    s16 x;
+    s16 y;
+    s16 z;
+};
+
+/**
  * Structure of data stored in HID shared memory
  */
 struct SharedMem {
@@ -112,6 +130,46 @@ struct SharedMem {
 
         std::array<TouchDataEntry, 8> entries; ///< Last 8 touch entries, in pixel coordinates
     } touch;
+
+    /// Accelerometer data
+    struct {
+        s64 index_reset_ticks; ///< CPU tick count for when HID module updated entry index 0
+        s64 index_reset_ticks_previous; ///< Previous `index_reset_ticks`
+        u32 index; ///< Index of the last updated accelerometer entry
+
+        INSERT_PADDING_WORDS(0x1);
+
+        AccelerometerDataEntry raw_entry;
+        INSERT_PADDING_BYTES(2);
+
+        std::array<AccelerometerDataEntry, 8> entries;
+    } accelerometer;
+
+    /// Gyroscope data
+    struct {
+        s64 index_reset_ticks; ///< CPU tick count for when HID module updated entry index 0
+        s64 index_reset_ticks_previous; ///< Previous `index_reset_ticks`
+        u32 index; ///< Index of the last updated accelerometer entry
+
+        INSERT_PADDING_WORDS(0x1);
+
+        GyroscopeDataEntry raw_entry;
+        INSERT_PADDING_BYTES(2);
+
+        std::array<GyroscopeDataEntry, 32> entries;
+    } gyroscope;
+};
+
+/**
+ * Structure of calibrate params that GetGyroscopeLowCalibrateParam returns
+ */
+struct GyroscopeCalibrateParam {
+    struct {
+        // TODO (wwylele): figure out the exact meaning of these params
+        s16 zero_point;
+        s16 positive_unit_point;
+        s16 negative_unit_point;
+    } x, y, z;
 };
 
 // TODO: MSVC does not support using offsetof() on non-static data members even though this
@@ -222,6 +280,26 @@ void DisableGyroscopeLow(Interface* self);
  */
 void GetSoundVolume(Interface* self);
 
+/**
+ * HID::GetGyroscopeLowRawToDpsCoefficient service function
+ *  Inputs:
+ *      None
+ *  Outputs:
+ *      1 : Result of function, 0 on success, otherwise error code
+ *      2 : float output value
+ */
+void GetGyroscopeLowRawToDpsCoefficient(Service::Interface* self);
+
+/**
+ * HID::GetGyroscopeLowCalibrateParam service function
+ *  Inputs:
+ *      None
+ *  Outputs:
+ *      1 : Result of function, 0 on success, otherwise error code
+ *      2~6 (18 bytes) : struct GyroscopeCalibrateParam
+ */
+void GetGyroscopeLowCalibrateParam(Service::Interface* self);
+
 /// Checks for user input updates
 void Update();
 
diff --git a/src/core/hle/service/hid/hid_spvr.cpp b/src/core/hle/service/hid/hid_spvr.cpp
index c50f597eb..046e65b11 100644
--- a/src/core/hle/service/hid/hid_spvr.cpp
+++ b/src/core/hle/service/hid/hid_spvr.cpp
@@ -9,16 +9,16 @@ namespace Service {
 namespace HID {
 
 const Interface::FunctionInfo FunctionTable[] = {
-    {0x000A0000, GetIPCHandles,              "GetIPCHandles"},
-    {0x000B0000, nullptr,                    "StartAnalogStickCalibration"},
-    {0x000E0000, nullptr,                    "GetAnalogStickCalibrateParam"},
-    {0x00110000, EnableAccelerometer,        "EnableAccelerometer"},
-    {0x00120000, DisableAccelerometer,       "DisableAccelerometer"},
-    {0x00130000, EnableGyroscopeLow,         "EnableGyroscopeLow"},
-    {0x00140000, DisableGyroscopeLow,        "DisableGyroscopeLow"},
-    {0x00150000, nullptr,                    "GetGyroscopeLowRawToDpsCoefficient"},
-    {0x00160000, nullptr,                    "GetGyroscopeLowCalibrateParam"},
-    {0x00170000, GetSoundVolume,             "GetSoundVolume"},
+    {0x000A0000, GetIPCHandles,                      "GetIPCHandles"},
+    {0x000B0000, nullptr,                            "StartAnalogStickCalibration"},
+    {0x000E0000, nullptr,                            "GetAnalogStickCalibrateParam"},
+    {0x00110000, EnableAccelerometer,                "EnableAccelerometer"},
+    {0x00120000, DisableAccelerometer,               "DisableAccelerometer"},
+    {0x00130000, EnableGyroscopeLow,                 "EnableGyroscopeLow"},
+    {0x00140000, DisableGyroscopeLow,                "DisableGyroscopeLow"},
+    {0x00150000, GetGyroscopeLowRawToDpsCoefficient, "GetGyroscopeLowRawToDpsCoefficient"},
+    {0x00160000, GetGyroscopeLowCalibrateParam,      "GetGyroscopeLowCalibrateParam"},
+    {0x00170000, GetSoundVolume,                     "GetSoundVolume"},
 };
 
 HID_SPVR_Interface::HID_SPVR_Interface() {
diff --git a/src/core/hle/service/hid/hid_user.cpp b/src/core/hle/service/hid/hid_user.cpp
index bbdde2abb..bb157b83d 100644
--- a/src/core/hle/service/hid/hid_user.cpp
+++ b/src/core/hle/service/hid/hid_user.cpp
@@ -9,16 +9,16 @@ namespace Service {
 namespace HID {
 
 const Interface::FunctionInfo FunctionTable[] = {
-    {0x000A0000, GetIPCHandles,             "GetIPCHandles"},
-    {0x000B0000, nullptr,                   "StartAnalogStickCalibration"},
-    {0x000E0000, nullptr,                   "GetAnalogStickCalibrateParam"},
-    {0x00110000, EnableAccelerometer,       "EnableAccelerometer"},
-    {0x00120000, DisableAccelerometer,      "DisableAccelerometer"},
-    {0x00130000, EnableGyroscopeLow,        "EnableGyroscopeLow"},
-    {0x00140000, DisableGyroscopeLow,       "DisableGyroscopeLow"},
-    {0x00150000, nullptr,                   "GetGyroscopeLowRawToDpsCoefficient"},
-    {0x00160000, nullptr,                   "GetGyroscopeLowCalibrateParam"},
-    {0x00170000, GetSoundVolume,            "GetSoundVolume"},
+    {0x000A0000, GetIPCHandles,                      "GetIPCHandles"},
+    {0x000B0000, nullptr,                            "StartAnalogStickCalibration"},
+    {0x000E0000, nullptr,                            "GetAnalogStickCalibrateParam"},
+    {0x00110000, EnableAccelerometer,                "EnableAccelerometer"},
+    {0x00120000, DisableAccelerometer,               "DisableAccelerometer"},
+    {0x00130000, EnableGyroscopeLow,                 "EnableGyroscopeLow"},
+    {0x00140000, DisableGyroscopeLow,                "DisableGyroscopeLow"},
+    {0x00150000, GetGyroscopeLowRawToDpsCoefficient, "GetGyroscopeLowRawToDpsCoefficient"},
+    {0x00160000, GetGyroscopeLowCalibrateParam,      "GetGyroscopeLowCalibrateParam"},
+    {0x00170000, GetSoundVolume,                     "GetSoundVolume"},
 };
 
 HID_U_Interface::HID_U_Interface() {
diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp
index 35b648409..7844d2330 100644
--- a/src/core/hle/service/service.cpp
+++ b/src/core/hle/service/service.cpp
@@ -9,6 +9,7 @@
 #include "core/hle/service/ac_u.h"
 #include "core/hle/service/act_u.h"
 #include "core/hle/service/csnd_snd.h"
+#include "core/hle/service/dlp_srvr.h"
 #include "core/hle/service/dsp_dsp.h"
 #include "core/hle/service/err_f.h"
 #include "core/hle/service/gsp_gpu.h"
@@ -121,6 +122,7 @@ void Init() {
     AddService(new AC_U::Interface);
     AddService(new ACT_U::Interface);
     AddService(new CSND_SND::Interface);
+    AddService(new DLP_SRVR::Interface);
     AddService(new DSP_DSP::Interface);
     AddService(new GSP_GPU::Interface);
     AddService(new GSP_LCD::Interface);
diff --git a/src/core/hle/service/soc_u.cpp b/src/core/hle/service/soc_u.cpp
index b52e52d4a..ff0af8f12 100644
--- a/src/core/hle/service/soc_u.cpp
+++ b/src/core/hle/service/soc_u.cpp
@@ -5,6 +5,7 @@
 #include <algorithm>
 #include <cstring>
 #include <unordered_map>
+#include <vector>
 
 #include "common/assert.h"
 #include "common/bit_field.h"
@@ -593,17 +594,13 @@ static void Poll(Service::Interface* self) {
 
     // The 3ds_pollfd and the pollfd structures may be different (Windows/Linux have different sizes)
     // so we have to copy the data
-    pollfd* platform_pollfd = new pollfd[nfds];
-    for (unsigned current_fds = 0; current_fds < nfds; ++current_fds)
-        platform_pollfd[current_fds] = CTRPollFD::ToPlatform(input_fds[current_fds]);
+    std::vector<pollfd> platform_pollfd(nfds);
+    std::transform(input_fds, input_fds + nfds, platform_pollfd.begin(), CTRPollFD::ToPlatform);
 
-    int ret = ::poll(platform_pollfd, nfds, timeout);
+    const int ret = ::poll(platform_pollfd.data(), nfds, timeout);
 
     // Now update the output pollfd structure
-    for (unsigned current_fds = 0; current_fds < nfds; ++current_fds)
-        output_fds[current_fds] = CTRPollFD::FromPlatform(platform_pollfd[current_fds]);
-
-    delete[] platform_pollfd;
+    std::transform(platform_pollfd.begin(), platform_pollfd.end(), output_fds, CTRPollFD::FromPlatform);
 
     int result = 0;
     if (ret == SOCKET_ERROR_VALUE)
diff --git a/src/core/loader/loader.h b/src/core/loader/loader.h
index a7f2715ba..84a4ce5fc 100644
--- a/src/core/loader/loader.h
+++ b/src/core/loader/loader.h
@@ -74,7 +74,7 @@ enum class ResultStatus {
     ErrorEncrypted,
 };
 
-static inline u32 MakeMagic(char a, char b, char c, char d) {
+constexpr u32 MakeMagic(char a, char b, char c, char d) {
     return a | b << 8 | c << 16 | d << 24;
 }
 
diff --git a/src/video_core/command_processor.cpp b/src/video_core/command_processor.cpp
index 4b59984ad..028b59348 100644
--- a/src/video_core/command_processor.cpp
+++ b/src/video_core/command_processor.cpp
@@ -75,12 +75,17 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
             GSP_GPU::SignalInterrupt(GSP_GPU::InterruptId::P3D);
             break;
 
+        case PICA_REG_INDEX_WORKAROUND(triangle_topology, 0x25E):
+            g_state.primitive_assembler.Reconfigure(regs.triangle_topology);
+            break;
+
+        case PICA_REG_INDEX_WORKAROUND(restart_primitive, 0x25F):
+            g_state.primitive_assembler.Reset();
+            break;
+
         case PICA_REG_INDEX_WORKAROUND(vs_default_attributes_setup.index, 0x232):
-            if (regs.vs_default_attributes_setup.index == 15) {
-                // Reset immediate primitive state
-                g_state.immediate.primitive_assembler.Reconfigure(regs.triangle_topology);
-                g_state.immediate.attribute_id = 0;
-            }
+            g_state.immediate.current_attribute = 0;
+            default_attr_counter = 0;
             break;
 
         // Load default vertex input attributes
@@ -105,7 +110,7 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
                     break;
                 }
 
-                Math::Vec4<float24>& attribute = g_state.vs.default_attributes[setup.index];
+                Math::Vec4<float24> attribute;
 
                 // NOTE: The destination component order indeed is "backwards"
                 attribute.w = float24::FromRaw(default_attr_write_buffer[0] >> 8);
@@ -119,26 +124,29 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
 
                 // TODO: Verify that this actually modifies the register!
                 if (setup.index < 15) {
+                    g_state.vs.default_attributes[setup.index] = attribute;
                     setup.index++;
                 } else {
                     // Put each attribute into an immediate input buffer.
                     // When all specified immediate attributes are present, the Vertex Shader is invoked and everything is
                     // sent to the primitive assembler.
 
-                    auto& immediate_input = g_state.immediate.input;
-                    auto& immediate_attribute_id = g_state.immediate.attribute_id;
-                    const auto& attribute_config = regs.vertex_attributes;
+                    auto& immediate_input = g_state.immediate.input_vertex;
+                    auto& immediate_attribute_id = g_state.immediate.current_attribute;
 
                     immediate_input.attr[immediate_attribute_id++] = attribute;
 
-                    if (immediate_attribute_id >= attribute_config.GetNumTotalAttributes()) {
+                    if (immediate_attribute_id >= regs.vs.num_input_attributes+1) {
                         immediate_attribute_id = 0;
 
                         Shader::UnitState<false> shader_unit;
                         Shader::Setup(shader_unit);
 
+                        if (g_debug_context)
+                            g_debug_context->OnEvent(DebugContext::Event::VertexLoaded, static_cast<void*>(&immediate_input));
+
                         // Send to vertex shader
-                        Shader::OutputVertex output = Shader::Run(shader_unit, immediate_input, attribute_config.GetNumTotalAttributes());
+                        Shader::OutputVertex output = Shader::Run(shader_unit, immediate_input, regs.vs.num_input_attributes+1);
 
                         // Send to renderer
                         using Pica::Shader::OutputVertex;
@@ -146,7 +154,7 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
                             VideoCore::g_renderer->Rasterizer()->AddTriangle(v0, v1, v2);
                         };
 
-                        g_state.immediate.primitive_assembler.SubmitVertex(output, AddTriangle);
+                        g_state.primitive_assembler.SubmitVertex(output, AddTriangle);
                     }
                 }
             }
@@ -154,9 +162,13 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
         }
 
         case PICA_REG_INDEX(gpu_mode):
-            if (regs.gpu_mode == Regs::GPUMode::Configuring && regs.vs_default_attributes_setup.index == 15) {
+            if (regs.gpu_mode == Regs::GPUMode::Configuring) {
                 // Draw immediate mode triangles when GPU Mode is set to GPUMode::Configuring
                 VideoCore::g_renderer->Rasterizer()->DrawTriangles();
+
+                if (g_debug_context) {
+                    g_debug_context->OnEvent(DebugContext::Event::FinishedPrimitiveBatch, nullptr);
+                }
             }
             break;
 
@@ -241,7 +253,7 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
             DebugUtils::GeometryDumper geometry_dumper;
             PrimitiveAssembler<DebugUtils::GeometryDumper::Vertex> dumping_primitive_assembler(regs.triangle_topology.Value());
 #endif
-            PrimitiveAssembler<Shader::OutputVertex> primitive_assembler(regs.triangle_topology.Value());
+            PrimitiveAssembler<Shader::OutputVertex>& primitive_assembler = g_state.primitive_assembler;
 
             if (g_debug_context) {
                 for (int i = 0; i < 3; ++i) {
@@ -412,16 +424,10 @@ static void WritePicaReg(u32 id, u32 value, u32 mask) {
                                                           range.second, range.first);
             }
 
-            VideoCore::g_renderer->Rasterizer()->DrawTriangles();
-
 #if PICA_DUMP_GEOMETRY
             geometry_dumper.Dump();
 #endif
 
-            if (g_debug_context) {
-                g_debug_context->OnEvent(DebugContext::Event::FinishedPrimitiveBatch, nullptr);
-            }
-
             break;
         }
 
diff --git a/src/video_core/pica.cpp b/src/video_core/pica.cpp
index 32ad72674..ccbaf071b 100644
--- a/src/video_core/pica.cpp
+++ b/src/video_core/pica.cpp
@@ -493,12 +493,25 @@ std::string Regs::GetCommandName(int index) {
 }
 
 void Init() {
+    g_state.Reset();
 }
 
 void Shutdown() {
     Shader::Shutdown();
+}
+
+template <typename T>
+void Zero(T& o) {
+    memset(&o, 0, sizeof(o));
+}
 
-    memset(&g_state, 0, sizeof(State));
+void State::Reset() {
+    Zero(regs);
+    Zero(vs);
+    Zero(gs);
+    Zero(cmd_list);
+    Zero(immediate);
+    primitive_assembler.Reconfigure(Regs::TriangleTopology::List);
 }
 
 }
diff --git a/src/video_core/pica.h b/src/video_core/pica.h
index 337cff8ce..16f9e4006 100644
--- a/src/video_core/pica.h
+++ b/src/video_core/pica.h
@@ -71,7 +71,7 @@ struct Regs {
     BitField<0, 24, u32> viewport_depth_range; // float24
     BitField<0, 24, u32> viewport_depth_far_plane; // float24
 
-    INSERT_PADDING_WORDS(0x1);
+    BitField<0, 3, u32> vs_output_total;
 
     union VSOutputAttributes {
         // Maps components of output vertex attributes to semantics
@@ -1123,7 +1123,12 @@ struct Regs {
             BitField<24, 8, u32> w;
         } int_uniforms[4];
 
-        INSERT_PADDING_WORDS(0x5);
+        INSERT_PADDING_WORDS(0x4);
+
+        union {
+            // Number of input attributes to shader unit - 1
+            BitField<0, 4, u32> num_input_attributes;
+        };
 
         // Offset to shader program entry point (in words)
         BitField<0, 16, u32> main_offset;
@@ -1157,8 +1162,10 @@ struct Regs {
             }
         } input_register_map;
 
-        // OUTMAP_MASK, 0x28E, CODETRANSFER_END
-        INSERT_PADDING_WORDS(0x3);
+        BitField<0, 16, u32> output_mask;
+
+        // 0x28E, CODETRANSFER_END
+        INSERT_PADDING_WORDS(0x2);
 
         struct {
             enum Format : u32
diff --git a/src/video_core/pica_state.h b/src/video_core/pica_state.h
index c7616bc55..323290054 100644
--- a/src/video_core/pica_state.h
+++ b/src/video_core/pica_state.h
@@ -12,6 +12,8 @@ namespace Pica {
 
 /// Struct used to describe current Pica state
 struct State {
+    void Reset();
+
     /// Pica registers
     Regs regs;
 
@@ -46,13 +48,14 @@ struct State {
 
     /// Struct used to describe immediate mode rendering state
     struct ImmediateModeState {
-        Shader::InputVertex input;
-        // This is constructed with a dummy triangle topology
-        PrimitiveAssembler<Shader::OutputVertex> primitive_assembler;
-        int attribute_id = 0;
-
-        ImmediateModeState() : primitive_assembler(Regs::TriangleTopology::List) {}
+        // Used to buffer partial vertices for immediate-mode rendering.
+        Shader::InputVertex input_vertex;
+        // Index of the next attribute to be loaded into `input_vertex`.
+        int current_attribute = 0;
     } immediate;
+
+    // This is constructed with a dummy triangle topology
+    PrimitiveAssembler<Shader::OutputVertex> primitive_assembler;
 };
 
 extern State g_state; ///< Current Pica state
diff --git a/src/video_core/primitive_assembly.h b/src/video_core/primitive_assembly.h
index cc6e5fde5..9396b4c85 100644
--- a/src/video_core/primitive_assembly.h
+++ b/src/video_core/primitive_assembly.h
@@ -20,7 +20,7 @@ struct PrimitiveAssembler {
                                                VertexType& v1,
                                                VertexType& v2)>;
 
-    PrimitiveAssembler(Regs::TriangleTopology topology);
+    PrimitiveAssembler(Regs::TriangleTopology topology = Regs::TriangleTopology::List);
 
     /*
      * Queues a vertex, builds primitives from the vertex queue according to the given
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 5d91aac00..cc6b3aeab 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -189,6 +189,9 @@ void RasterizerOpenGL::AddTriangle(const Pica::Shader::OutputVertex& v0,
 }
 
 void RasterizerOpenGL::DrawTriangles() {
+    if (vertex_batch.empty())
+        return;
+
     SyncFramebuffer();
     SyncDrawState();
 
@@ -804,6 +807,9 @@ void RasterizerOpenGL::SyncFramebuffer() {
 
         ReloadDepthBuffer();
     }
+
+    ASSERT_MSG(glCheckFramebufferStatus(GL_FRAMEBUFFER) == GL_FRAMEBUFFER_COMPLETE,
+               "OpenGL rasterizer framebuffer setup failed, status %X", glCheckFramebufferStatus(GL_FRAMEBUFFER));
 }
 
 void RasterizerOpenGL::SyncCullMode() {
diff --git a/src/video_core/renderer_opengl/pica_to_gl.h b/src/video_core/renderer_opengl/pica_to_gl.h
index 3d6c4e9e5..fd3617d77 100644
--- a/src/video_core/renderer_opengl/pica_to_gl.h
+++ b/src/video_core/renderer_opengl/pica_to_gl.h
@@ -22,7 +22,7 @@ inline GLenum TextureFilterMode(Pica::Regs::TextureConfig::TextureFilter mode) {
     };
 
     // Range check table for input
-    if (mode >= ARRAY_SIZE(filter_mode_table)) {
+    if (static_cast<size_t>(mode) >= ARRAY_SIZE(filter_mode_table)) {
         LOG_CRITICAL(Render_OpenGL, "Unknown texture filtering mode %d", mode);
         UNREACHABLE();
 
@@ -51,7 +51,7 @@ inline GLenum WrapMode(Pica::Regs::TextureConfig::WrapMode mode) {
     };
 
     // Range check table for input
-    if (mode >= ARRAY_SIZE(wrap_mode_table)) {
+    if (static_cast<size_t>(mode) >= ARRAY_SIZE(wrap_mode_table)) {
         LOG_CRITICAL(Render_OpenGL, "Unknown texture wrap mode %d", mode);
         UNREACHABLE();
 
@@ -91,7 +91,7 @@ inline GLenum BlendFunc(Pica::Regs::BlendFactor factor) {
     };
 
     // Range check table for input
-    if ((unsigned)factor >= ARRAY_SIZE(blend_func_table)) {
+    if (static_cast<size_t>(factor) >= ARRAY_SIZE(blend_func_table)) {
         LOG_CRITICAL(Render_OpenGL, "Unknown blend factor %d", factor);
         UNREACHABLE();
 
@@ -122,7 +122,7 @@ inline GLenum LogicOp(Pica::Regs::LogicOp op) {
     };
 
     // Range check table for input
-    if ((unsigned)op >= ARRAY_SIZE(logic_op_table)) {
+    if (static_cast<size_t>(op) >= ARRAY_SIZE(logic_op_table)) {
         LOG_CRITICAL(Render_OpenGL, "Unknown logic op %d", op);
         UNREACHABLE();
 
@@ -145,7 +145,7 @@ inline GLenum CompareFunc(Pica::Regs::CompareFunc func) {
     };
 
     // Range check table for input
-    if ((unsigned)func >= ARRAY_SIZE(compare_func_table)) {
+    if (static_cast<size_t>(func) >= ARRAY_SIZE(compare_func_table)) {
         LOG_CRITICAL(Render_OpenGL, "Unknown compare function %d", func);
         UNREACHABLE();
 
@@ -168,7 +168,7 @@ inline GLenum StencilOp(Pica::Regs::StencilAction action) {
     };
 
     // Range check table for input
-    if ((unsigned)action >= ARRAY_SIZE(stencil_op_table)) {
+    if (static_cast<size_t>(action) >= ARRAY_SIZE(stencil_op_table)) {
         LOG_CRITICAL(Render_OpenGL, "Unknown stencil op %d", action);
         UNREACHABLE();
 
diff --git a/src/video_core/shader/shader.cpp b/src/video_core/shader/shader.cpp
index 045e7d5bd..78d295c76 100644
--- a/src/video_core/shader/shader.cpp
+++ b/src/video_core/shader/shader.cpp
@@ -120,15 +120,23 @@ OutputVertex Run(UnitState<false>& state, const InputVertex& input, int num_attr
     OutputVertex ret;
     // TODO(neobrain): Under some circumstances, up to 16 attributes may be output. We need to
     // figure out what those circumstances are and enable the remaining outputs then.
-    for (int i = 0; i < 7; ++i) {
-        const auto& output_register_map = g_state.regs.vs_output_attributes[i]; // TODO: Don't hardcode VS here
+    unsigned index = 0;
+    for (unsigned i = 0; i < 7; ++i) {
+
+        if (index >= g_state.regs.vs_output_total)
+            break;
+
+        if ((g_state.regs.vs.output_mask & (1 << i)) == 0)
+            continue;
+
+        const auto& output_register_map = g_state.regs.vs_output_attributes[index]; // TODO: Don't hardcode VS here
 
         u32 semantics[4] = {
             output_register_map.map_x, output_register_map.map_y,
             output_register_map.map_z, output_register_map.map_w
         };
 
-        for (int comp = 0; comp < 4; ++comp) {
+        for (unsigned comp = 0; comp < 4; ++comp) {
             float24* out = ((float24*)&ret) + semantics[comp];
             if (semantics[comp] != Regs::VSOutputAttributes::INVALID) {
                 *out = state.registers.output[i][comp];
@@ -138,10 +146,12 @@ OutputVertex Run(UnitState<false>& state, const InputVertex& input, int num_attr
                 memset(out, 0, sizeof(*out));
             }
         }
+
+        index++;
     }
 
     // The hardware takes the absolute and saturates vertex colors like this, *before* doing interpolation
-    for (int i = 0; i < 4; ++i) {
+    for (unsigned i = 0; i < 4; ++i) {
         ret.color[i] = float24::FromFloat32(
             std::fmin(std::fabs(ret.color[i].ToFloat32()), 1.0f));
     }