24 files changed, 246 insertions, 200 deletions

diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index c8c8e3884..906a4bc7d 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -10,16 +10,16 @@ Citra is a brand new project, so we have a great opportunity to keep things clea
 
 ### Naming Rules
 * Functions
- * CamelCase, "_" may also be used for clarity (e.g. ARM_InitCore)
+ * PascalCase, "_" may also be used for clarity (e.g. ARM_InitCore)
 * Variables
  * lower_case_underscored
  * Prefix "g_" if global
 * Classes
- * CamelCase, "_" may also be used for clarity (e.g. OGL_VideoInterface)
+ * PascalCase, "_" may also be used for clarity (e.g. OGL_VideoInterface)
 * Files/Folders
  * lower_case_underscored
 * Namespaces
- * CamelCase, "_" may also be used for clarity (e.g. ARM_InitCore)
+ * PascalCase, "_" may also be used for clarity (e.g. ARM_InitCore)
 
 ### Indentation/Whitespace Style
 Follow the indentation/whitespace style shown below. Do not use tabs, use 4-spaces instead.
diff --git a/appveyor.yml b/appveyor.yml
index 46362d1b4..8b420f20c 100644
--- a/appveyor.yml
+++ b/appveyor.yml
@@ -33,12 +33,15 @@ after_build:
             $GITREV = $(git show -s --format='%h')
             # Where are these spaces coming from? Regardless, let's remove them
             $BUILD_NAME = "citra-${GITDATE}-${GITREV}-windows-amd64.7z" -replace " ",""
+            $BUILD_NAME_NOQT = "citra-noqt-${GITDATE}-${GITREV}-windows-amd64.7z" -replace " ",""
             # Zip up the build folder
             7z a $BUILD_NAME .\build\bin\release\*
+            # Do a second archive with only the binaries
+            7z a $BUILD_NAME_NOQT .\build\bin\release\*.exe
 
             # Download winscp
-            Invoke-WebRequest "http://hivelocity.dl.sourceforge.net/project/winscp/WinSCP/5.7/winscp570.zip" -OutFile "winscp570.zip"
-            7z e -y winscp570.zip
+            Invoke-WebRequest "https://cdn.winscp.net/files/winscp573.zip?secure=yE1jHlRoBZTA0htWnu1OVQ==,1434405422" -OutFile "winscp573.zip"
+            7z e -y winscp573.zip
 
             # Upload to server
             .\WinSCP.com /command `
@@ -46,5 +49,6 @@ after_build:
                 "option confirm off" `
                 "open sftp://citra-builds:${env:BUILD_PASSWORD}@builds.citra-emu.org -hostkey=*" `
                 "put $BUILD_NAME /citra/nightly/windows-amd64/" `
+                "put $BUILD_NAME_NOQT /citra/nightly/windows-noqt-amd64/" `
                 "exit"
           }
diff --git a/src/citra_qt/CMakeLists.txt b/src/citra_qt/CMakeLists.txt
index c05779380..c2d1ad240 100644
--- a/src/citra_qt/CMakeLists.txt
+++ b/src/citra_qt/CMakeLists.txt
@@ -91,10 +91,10 @@ if (Qt5_FOUND AND MSVC)
     set(PLATFORMS ${DLL_DEST}platforms/)
 
     # windows commandline expects the / to be \ so switch them
-    string(REPLACE "/" "\\" Qt5_DLL_DIR ${Qt5_DLL_DIR})
-    string(REPLACE "/" "\\" Qt5_PLATFORMS_DIR ${Qt5_PLATFORMS_DIR})
-    string(REPLACE "/" "\\" DLL_DEST ${DLL_DEST})
-    string(REPLACE "/" "\\" PLATFORMS ${PLATFORMS})
+    string(REPLACE "/" "\\\\" Qt5_DLL_DIR ${Qt5_DLL_DIR})
+    string(REPLACE "/" "\\\\" Qt5_PLATFORMS_DIR ${Qt5_PLATFORMS_DIR})
+    string(REPLACE "/" "\\\\" DLL_DEST ${DLL_DEST})
+    string(REPLACE "/" "\\\\" PLATFORMS ${PLATFORMS})
 
     # /NJH /NJS /NDL /NFL /NC /NS /NP - Silence any output
     # cmake adds an extra check for command success which doesn't work too well with robocopy
diff --git a/src/citra_qt/debugger/callstack.ui b/src/citra_qt/debugger/callstack.ui
index b0e31120f..248ea3dd7 100644
--- a/src/citra_qt/debugger/callstack.ui
+++ b/src/citra_qt/debugger/callstack.ui
@@ -17,6 +17,9 @@
    <layout class="QVBoxLayout" name="verticalLayout">
     <item>
      <widget class="QTreeView" name="treeView">
+      <property name="editTriggers">
+       <set>QAbstractItemView::NoEditTriggers</set>
+      </property>
       <property name="alternatingRowColors">
        <bool>true</bool>
       </property>
diff --git a/src/common/math_util.h b/src/common/math_util.h
index 4b0910741..d44b06e74 100644
--- a/src/common/math_util.h
+++ b/src/common/math_util.h
@@ -12,7 +12,7 @@ namespace MathUtil
 {
 
 inline bool IntervalsIntersect(unsigned start0, unsigned length0, unsigned start1, unsigned length1) {
-    return (std::max(start0, start1) <= std::min(start0 + length0, start1 + length1));
+    return (std::max(start0, start1) < std::min(start0 + length0, start1 + length1));
 }
 
 template<typename T>
diff --git a/src/core/arm/skyeye_common/arm_regformat.h b/src/core/arm/skyeye_common/arm_regformat.h
index 6c89774eb..a92effbb4 100644
--- a/src/core/arm/skyeye_common/arm_regformat.h
+++ b/src/core/arm/skyeye_common/arm_regformat.h
@@ -59,6 +59,8 @@ enum {
     VFP_FPSID,
     VFP_FPSCR,
     VFP_FPEXC,
+    VFP_MVFR0,
+    VFP_MVFR1,
 
     // Not an actual register.
     // All VFP system registers should be defined above this.
diff --git a/src/core/arm/skyeye_common/vfp/vfp.cpp b/src/core/arm/skyeye_common/vfp/vfp.cpp
index b88d47750..571d6c2f2 100644
--- a/src/core/arm/skyeye_common/vfp/vfp.cpp
+++ b/src/core/arm/skyeye_common/vfp/vfp.cpp
@@ -33,6 +33,10 @@ unsigned VFPInit(ARMul_State* state)
     state->VFP[VFP_FPEXC] = 0;
     state->VFP[VFP_FPSCR] = 0;
 
+    // ARM11 MPCore feature register values.
+    state->VFP[VFP_MVFR0] = 0x11111111;
+    state->VFP[VFP_MVFR1] = 0;
+
     return 0;
 }
 
diff --git a/src/core/arm/skyeye_common/vfp/vfp.h b/src/core/arm/skyeye_common/vfp/vfp.h
index 727350f14..acefae9bb 100644
--- a/src/core/arm/skyeye_common/vfp/vfp.h
+++ b/src/core/arm/skyeye_common/vfp/vfp.h
@@ -22,7 +22,6 @@
 
 #include "core/arm/skyeye_common/vfp/vfp_helper.h" /* for references to cdp SoftFloat functions */
 
-#define VFP_DEBUG_UNIMPLEMENTED(x) LOG_ERROR(Core_ARM11, "in func %s, " #x " unimplemented\n", __FUNCTION__); exit(-1);
 #define VFP_DEBUG_UNTESTED(x) LOG_TRACE(Core_ARM11, "in func %s, " #x " untested\n", __FUNCTION__);
 #define CHECK_VFP_ENABLED
 #define CHECK_VFP_CDP_RET vfp_raise_exceptions(cpu, ret, inst_cream->instr, cpu->VFP[VFP_FPSCR]);
diff --git a/src/core/arm/skyeye_common/vfp/vfpinstr.cpp b/src/core/arm/skyeye_common/vfp/vfpinstr.cpp
index 3ed918a93..67fe63aa4 100644
--- a/src/core/arm/skyeye_common/vfp/vfpinstr.cpp
+++ b/src/core/arm/skyeye_common/vfp/vfpinstr.cpp
@@ -1068,10 +1068,12 @@ static ARM_INST_PTR INTERPRETER_TRANSLATE(vmovbrc)(unsigned int inst, int index)
 #ifdef VFP_INTERPRETER_IMPL
 VMOVBRC_INST:
 {
-    if ((inst_base->cond == 0xe) || CondPassed(cpu, inst_base->cond)) {
+    if (inst_base->cond == 0xE || CondPassed(cpu, inst_base->cond)) {
         CHECK_VFP_ENABLED;
 
-        VFP_DEBUG_UNIMPLEMENTED(VMOVBRC);
+        vmovbrc_inst* const inst_cream = (vmovbrc_inst*)inst_base->component;
+
+        cpu->ExtReg[(2 * inst_cream->d) + inst_cream->index] = cpu->Reg[inst_cream->t];
     }
     cpu->Reg[15] += GET_INST_SIZE(cpu);
     INC_PC(sizeof(vmovbrc_inst));
@@ -1139,12 +1141,10 @@ VMRS_INST:
                 cpu->Reg[inst_cream->Rt] = cpu->VFP[VFP_FPSID];
                 break;
             case 6:
-                /* MVFR1, VFPv3 only ? */
-                LOG_TRACE(Core_ARM11, "\tr%d <= MVFR1 unimplemented\n", inst_cream->Rt);
+                cpu->Reg[inst_cream->Rt] = cpu->VFP[VFP_MVFR1];
                 break;
             case 7:
-                /* MVFR0, VFPv3 only? */
-                LOG_TRACE(Core_ARM11, "\tr%d <= MVFR0 unimplemented\n", inst_cream->Rt);
+                cpu->Reg[inst_cream->Rt] = cpu->VFP[VFP_MVFR0];
                 break;
             case 8:
                 cpu->Reg[inst_cream->Rt] = cpu->VFP[VFP_FPEXC];
@@ -1195,10 +1195,12 @@ static ARM_INST_PTR INTERPRETER_TRANSLATE(vmovbcr)(unsigned int inst, int index)
 #ifdef VFP_INTERPRETER_IMPL
 VMOVBCR_INST:
 {
-    if ((inst_base->cond == 0xe) || CondPassed(cpu, inst_base->cond)) {
+    if (inst_base->cond == 0xE || CondPassed(cpu, inst_base->cond)) {
         CHECK_VFP_ENABLED;
 
-        VFP_DEBUG_UNIMPLEMENTED(VMOVBCR);
+        vmovbcr_inst* const inst_cream = (vmovbcr_inst*) inst_base->component;
+
+        cpu->Reg[inst_cream->t] = cpu->ExtReg[(2 * inst_cream->d) + inst_cream->index];
     }
     cpu->Reg[15] += GET_INST_SIZE(cpu);
     INC_PC(sizeof(vmovbcr_inst));
diff --git a/src/core/hle/function_wrappers.h b/src/core/hle/function_wrappers.h
index 23c86a72d..5949cb470 100644
--- a/src/core/hle/function_wrappers.h
+++ b/src/core/hle/function_wrappers.h
@@ -9,11 +9,15 @@
 #include "core/arm/arm_interface.h"
 #include "core/memory.h"
 #include "core/hle/hle.h"
+#include "core/hle/result.h"
 
 namespace HLE {
 
 #define PARAM(n)    Core::g_app_core->GetReg(n)
 
+/// An invalid result code that is meant to be overwritten when a thread resumes from waiting
+static const ResultCode RESULT_INVALID(0xDEADC0DE);
+
 /**
  * HLE a function return from the current ARM11 userland process
  * @param res Result to return
@@ -57,8 +61,11 @@ template<ResultCode func(s32*, u32*, s32, bool, s64)> void Wrap() {
     s32 param_1 = 0;
     s32 retval = func(&param_1, (Handle*)Memory::GetPointer(PARAM(1)), (s32)PARAM(2),
         (PARAM(3) != 0), (((s64)PARAM(4) << 32) | PARAM(0))).raw;
-    Core::g_app_core->SetReg(1, (u32)param_1);
-    FuncReturn(retval);
+
+    if (retval != RESULT_INVALID.raw) {
+        Core::g_app_core->SetReg(1, (u32)param_1);
+        FuncReturn(retval);
+    }
 }
 
 template<ResultCode func(u32, u32, u32, u32, s64)> void Wrap() {
@@ -73,7 +80,11 @@ template<ResultCode func(u32*)> void Wrap(){
 }
 
 template<ResultCode func(u32, s64)> void Wrap() {
-    FuncReturn(func(PARAM(0), (((s64)PARAM(3) << 32) | PARAM(2))).raw);
+    s32 retval = func(PARAM(0), (((s64)PARAM(3) << 32) | PARAM(2))).raw;
+
+    if (retval != RESULT_INVALID.raw) {
+        FuncReturn(retval);
+    }
 }
 
 template<ResultCode func(void*, void*, u32)> void Wrap(){
diff --git a/src/core/hle/kernel/event.cpp b/src/core/hle/kernel/event.cpp
index e45deb1c6..f338f3266 100644
--- a/src/core/hle/kernel/event.cpp
+++ b/src/core/hle/kernel/event.cpp
@@ -41,10 +41,7 @@ void Event::Acquire() {
 
 void Event::Signal() {
     signaled = true;
-
     WakeupAllWaitingThreads();
-
-    HLE::Reschedule(__func__);
 }
 
 void Event::Clear() {
diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp
index 726e4d2ff..20e11da16 100644
--- a/src/core/hle/kernel/kernel.cpp
+++ b/src/core/hle/kernel/kernel.cpp
@@ -32,27 +32,13 @@ void WaitObject::RemoveWaitingThread(Thread* thread) {
         waiting_threads.erase(itr);
 }
 
-SharedPtr<Thread> WaitObject::WakeupNextThread() {
-    if (waiting_threads.empty())
-        return nullptr;
-
-    auto next_thread = std::move(waiting_threads.front());
-    waiting_threads.erase(waiting_threads.begin());
-
-    next_thread->ReleaseWaitObject(this);
-
-    return next_thread;
-}
-
 void WaitObject::WakeupAllWaitingThreads() {
-    auto waiting_threads_copy = waiting_threads;
+    for (auto thread : waiting_threads)
+        thread->ResumeFromWait();
 
-    // We use a copy because ReleaseWaitObject will remove the thread from this object's
-    // waiting_threads list
-    for (auto thread : waiting_threads_copy)
-        thread->ReleaseWaitObject(this);
+    waiting_threads.clear();
 
-    ASSERT_MSG(waiting_threads.empty(), "failed to awaken all waiting threads!");
+    HLE::Reschedule(__func__);
 }
 
 HandleTable::HandleTable() {
diff --git a/src/core/hle/kernel/kernel.h b/src/core/hle/kernel/kernel.h
index a5a0f4800..64595f758 100644
--- a/src/core/hle/kernel/kernel.h
+++ b/src/core/hle/kernel/kernel.h
@@ -140,12 +140,6 @@ public:
      */
     void RemoveWaitingThread(Thread* thread);
 
-    /**
-     * Wake up the next thread waiting on this object
-     * @return Pointer to the thread that was resumed, nullptr if no threads are waiting
-     */
-    SharedPtr<Thread> WakeupNextThread();
-
     /// Wake up all threads waiting on this object
     void WakeupAllWaitingThreads();
 
diff --git a/src/core/hle/kernel/mutex.cpp b/src/core/hle/kernel/mutex.cpp
index 6aa73df86..edb97d324 100644
--- a/src/core/hle/kernel/mutex.cpp
+++ b/src/core/hle/kernel/mutex.cpp
@@ -23,12 +23,7 @@ static void ResumeWaitingThread(Mutex* mutex) {
     // Reset mutex lock thread handle, nothing is waiting
     mutex->lock_count = 0;
     mutex->holding_thread = nullptr;
-
-    // Find the next waiting thread for the mutex...
-    auto next_thread = mutex->WakeupNextThread();
-    if (next_thread != nullptr) {
-        mutex->Acquire(next_thread);
-    }
+    mutex->WakeupAllWaitingThreads();
 }
 
 void ReleaseThreadMutexes(Thread* thread) {
@@ -94,8 +89,6 @@ void Mutex::Release() {
             ResumeWaitingThread(this);
         }
     }
-
-    HLE::Reschedule(__func__);
 }
 
 } // namespace
diff --git a/src/core/hle/kernel/semaphore.cpp b/src/core/hle/kernel/semaphore.cpp
index 96d61ed3a..4b359ed07 100644
--- a/src/core/hle/kernel/semaphore.cpp
+++ b/src/core/hle/kernel/semaphore.cpp
@@ -48,13 +48,7 @@ ResultVal<s32> Semaphore::Release(s32 release_count) {
     s32 previous_count = available_count;
     available_count += release_count;
 
-    // Notify some of the threads that the semaphore has been released
-    // stop once the semaphore is full again or there are no more waiting threads
-    while (!ShouldWait() && WakeupNextThread() != nullptr) {
-        Acquire();
-    }
-
-    HLE::Reschedule(__func__);
+    WakeupAllWaitingThreads();
 
     return MakeResult<s32>(previous_count);
 }
diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp
index 22c795ad4..4729a7fe0 100644
--- a/src/core/hle/kernel/thread.cpp
+++ b/src/core/hle/kernel/thread.cpp
@@ -13,6 +13,7 @@
 #include "common/thread_queue_list.h"
 
 #include "core/arm/arm_interface.h"
+#include "core/arm/skyeye_common/armdefs.h"
 #include "core/core.h"
 #include "core/core_timing.h"
 #include "core/hle/hle.h"
@@ -193,8 +194,22 @@ static void SwitchContext(Thread* new_thread) {
     if (new_thread) {
         DEBUG_ASSERT_MSG(new_thread->status == THREADSTATUS_READY, "Thread must be ready to become running.");
 
+        // Cancel any outstanding wakeup events for this thread
+        CoreTiming::UnscheduleEvent(ThreadWakeupEventType, new_thread->callback_handle);
+
         current_thread = new_thread;
 
+        // If the thread was waited by a svcWaitSynch call, step back PC by one instruction to rerun
+        // the SVC when the thread wakes up. This is necessary to ensure that the thread can acquire
+        // the requested wait object(s) before continuing.
+        if (new_thread->waitsynch_waited) {
+            // CPSR flag indicates CPU mode
+            bool thumb_mode = (new_thread->context.cpsr & TBIT) != 0;
+
+            // SVC instruction is 2 bytes for THUMB, 4 bytes for ARM
+            new_thread->context.pc -= thumb_mode ? 2 : 4;
+        }
+
         ready_queue.remove(new_thread->current_priority, new_thread);
         new_thread->status = THREADSTATUS_RUNNING;
 
@@ -243,6 +258,7 @@ void WaitCurrentThread_WaitSynchronization(std::vector<SharedPtr<WaitObject>> wa
     thread->wait_set_output = wait_set_output;
     thread->wait_all = wait_all;
     thread->wait_objects = std::move(wait_objects);
+    thread->waitsynch_waited = true;
     thread->status = THREADSTATUS_WAIT_SYNCH;
 }
 
@@ -268,6 +284,8 @@ static void ThreadWakeupCallback(u64 thread_handle, int cycles_late) {
         return;
     }
 
+    thread->waitsynch_waited = false;
+
     if (thread->status == THREADSTATUS_WAIT_SYNCH) {
         thread->SetWaitSynchronizationResult(ResultCode(ErrorDescription::Timeout, ErrorModule::OS,
                                                         ErrorSummary::StatusChanged, ErrorLevel::Info));
@@ -288,63 +306,20 @@ void Thread::WakeAfterDelay(s64 nanoseconds) {
     CoreTiming::ScheduleEvent(usToCycles(microseconds), ThreadWakeupEventType, callback_handle);
 }
 
-void Thread::ReleaseWaitObject(WaitObject* wait_object) {
-    if (status != THREADSTATUS_WAIT_SYNCH || wait_objects.empty()) {
-        LOG_CRITICAL(Kernel, "thread is not waiting on any objects!");
-        return;
-    }
-
-    // Remove this thread from the waiting object's thread list
-    wait_object->RemoveWaitingThread(this);
-
-    unsigned index = 0;
-    bool wait_all_failed = false; // Will be set to true if any object is unavailable
-
-    // Iterate through all waiting objects to check availability...
-    for (auto itr = wait_objects.begin(); itr != wait_objects.end(); ++itr) {
-        if ((*itr)->ShouldWait())
-            wait_all_failed = true;
-
-        // The output should be the last index of wait_object
-        if (*itr == wait_object)
-            index = itr - wait_objects.begin();
-    }
-
-    // If we are waiting on all objects...
-    if (wait_all) {
-        // Resume the thread only if all are available...
-        if (!wait_all_failed) {
-            SetWaitSynchronizationResult(RESULT_SUCCESS);
-            SetWaitSynchronizationOutput(-1);
-
-            ResumeFromWait();
-        }
-    } else {
-        // Otherwise, resume
-        SetWaitSynchronizationResult(RESULT_SUCCESS);
-
-        if (wait_set_output)
-            SetWaitSynchronizationOutput(index);
-
-        ResumeFromWait();
-    }
-}
-
 void Thread::ResumeFromWait() {
-    // Cancel any outstanding wakeup events for this thread
-    CoreTiming::UnscheduleEvent(ThreadWakeupEventType, callback_handle);
-
     switch (status) {
         case THREADSTATUS_WAIT_SYNCH:
-            // Remove this thread from all other WaitObjects
-            for (auto wait_object : wait_objects)
-                wait_object->RemoveWaitingThread(this);
-            break;
         case THREADSTATUS_WAIT_ARB:
         case THREADSTATUS_WAIT_SLEEP:
             break;
-        case THREADSTATUS_RUNNING:
+
         case THREADSTATUS_READY:
+            // If the thread is waiting on multiple wait objects, it might be awoken more than once
+            // before actually resuming. We can ignore subsequent wakeups if the thread status has
+            // already been set to THREADSTATUS_READY.
+            return;
+
+        case THREADSTATUS_RUNNING:
             DEBUG_ASSERT_MSG(false, "Thread with object id %u has already resumed.", GetObjectId());
             return;
         case THREADSTATUS_DEAD:
@@ -415,6 +390,7 @@ ResultVal<SharedPtr<Thread>> Thread::Create(std::string name, VAddr entry_point,
     thread->callback_handle = wakeup_callback_handle_table.Create(thread).MoveFrom();
     thread->owner_process = g_current_process;
     thread->tls_index = -1;
+    thread->waitsynch_waited = false;
 
     // Find the next available TLS index, and mark it as used
     auto& used_tls_slots = Kernel::g_current_process->used_tls_slots;
diff --git a/src/core/hle/kernel/thread.h b/src/core/hle/kernel/thread.h
index 2c65419c3..b8160bb2c 100644
--- a/src/core/hle/kernel/thread.h
+++ b/src/core/hle/kernel/thread.h
@@ -96,12 +96,6 @@ public:
     u32 GetThreadId() const { return thread_id; }
 
     /**
-     * Release an acquired wait object
-     * @param wait_object WaitObject to release
-     */
-    void ReleaseWaitObject(WaitObject* wait_object);
-
-    /**
      * Resumes a thread from waiting
      */
     void ResumeFromWait();
@@ -152,6 +146,8 @@ public:
 
     s32 tls_index; ///< Index of the Thread Local Storage of the thread
 
+    bool waitsynch_waited; ///< Set to true if the last svcWaitSynch call caused the thread to wait
+
     /// Mutexes currently held by this thread, which will be released when it exits.
     boost::container::flat_set<SharedPtr<Mutex>> held_mutexes;
 
@@ -163,12 +159,12 @@ public:
 
     std::string name;
 
+    /// Handle used as userdata to reference this object when inserting into the CoreTiming queue.
+    Handle callback_handle;
+
 private:
     Thread();
     ~Thread() override;
-
-    /// Handle used as userdata to reference this object when inserting into the CoreTiming queue.
-    Handle callback_handle;
 };
 
 /**
diff --git a/src/core/hle/svc.cpp b/src/core/hle/svc.cpp
index d1555c753..6cde4fc87 100644
--- a/src/core/hle/svc.cpp
+++ b/src/core/hle/svc.cpp
@@ -40,9 +40,6 @@ const ResultCode ERR_NOT_FOUND(ErrorDescription::NotFound, ErrorModule::Kernel,
 const ResultCode ERR_PORT_NAME_TOO_LONG(ErrorDescription(30), ErrorModule::OS,
         ErrorSummary::InvalidArgument, ErrorLevel::Usage); // 0xE0E0181E
 
-/// An invalid result code that is meant to be overwritten when a thread resumes from waiting
-const ResultCode RESULT_INVALID(0xDEADC0DE);
-
 enum ControlMemoryOperation {
     MEMORY_OPERATION_HEAP       = 0x00000003,
     MEMORY_OPERATION_GSP_HEAP   = 0x00010003,
@@ -143,6 +140,10 @@ static ResultCode CloseHandle(Handle handle) {
 /// Wait for a handle to synchronize, timeout after the specified nanoseconds
 static ResultCode WaitSynchronization1(Handle handle, s64 nano_seconds) {
     auto object = Kernel::g_handle_table.GetWaitObject(handle);
+    Kernel::Thread* thread = Kernel::GetCurrentThread();
+
+    thread->waitsynch_waited = false;
+
     if (object == nullptr)
         return ERR_INVALID_HANDLE;
 
@@ -154,14 +155,14 @@ static ResultCode WaitSynchronization1(Handle handle, s64 nano_seconds) {
     // Check for next thread to schedule
     if (object->ShouldWait()) {
 
-        object->AddWaitingThread(Kernel::GetCurrentThread());
+        object->AddWaitingThread(thread);
         Kernel::WaitCurrentThread_WaitSynchronization({ object }, false, false);
 
         // Create an event to wake the thread up after the specified nanosecond delay has passed
-        Kernel::GetCurrentThread()->WakeAfterDelay(nano_seconds);
+        thread->WakeAfterDelay(nano_seconds);
 
         // NOTE: output of this SVC will be set later depending on how the thread resumes
-        return RESULT_INVALID;
+        return HLE::RESULT_INVALID;
     }
 
     object->Acquire();
@@ -173,6 +174,9 @@ static ResultCode WaitSynchronization1(Handle handle, s64 nano_seconds) {
 static ResultCode WaitSynchronizationN(s32* out, Handle* handles, s32 handle_count, bool wait_all, s64 nano_seconds) {
     bool wait_thread = !wait_all;
     int handle_index = 0;
+    Kernel::Thread* thread = Kernel::GetCurrentThread();
+    bool was_waiting = thread->waitsynch_waited;
+    thread->waitsynch_waited = false;
 
     // Check if 'handles' is invalid
     if (handles == nullptr)
@@ -190,6 +194,9 @@ static ResultCode WaitSynchronizationN(s32* out, Handle* handles, s32 handle_cou
     // necessary
     if (handle_count != 0) {
         bool selected = false; // True once an object has been selected
+
+        Kernel::SharedPtr<Kernel::WaitObject> wait_object;
+
         for (int i = 0; i < handle_count; ++i) {
             auto object = Kernel::g_handle_table.GetWaitObject(handles[i]);
             if (object == nullptr)
@@ -204,10 +211,11 @@ static ResultCode WaitSynchronizationN(s32* out, Handle* handles, s32 handle_cou
                     wait_thread = true;
             } else {
                 // Do not wait on this object, check if this object should be selected...
-                if (!wait_all && !selected) {
+                if (!wait_all && (!selected || (wait_object == object && was_waiting))) {
                     // Do not wait the thread
                     wait_thread = false;
                     handle_index = i;
+                    wait_object = object;
                     selected = true;
                 }
             }
@@ -241,7 +249,7 @@ static ResultCode WaitSynchronizationN(s32* out, Handle* handles, s32 handle_cou
         Kernel::GetCurrentThread()->WakeAfterDelay(nano_seconds);
 
         // NOTE: output of this SVC will be set later depending on how the thread resumes
-        return RESULT_INVALID;
+        return HLE::RESULT_INVALID;
     }
 
     // Acquire objects if we did not wait...
@@ -261,7 +269,7 @@ static ResultCode WaitSynchronizationN(s32* out, Handle* handles, s32 handle_cou
 
     // TODO(bunnei): If 'wait_all' is true, this is probably wrong. However, real hardware does
     // not seem to set it to any meaningful value.
-    *out = wait_all ? 0 : handle_index;
+    *out = handle_count != 0 ? (wait_all ? -1 : handle_index) : 0;
 
     return RESULT_SUCCESS;
 }
diff --git a/src/core/hw/hw.cpp b/src/core/hw/hw.cpp
index f4906cc7e..c7006a498 100644
--- a/src/core/hw/hw.cpp
+++ b/src/core/hw/hw.cpp
@@ -18,7 +18,7 @@ inline void Read(T &var, const u32 addr) {
         GPU::Read(var, addr);
         break;
     case VADDR_LCD:
-        LCD::Write(var, addr);
+        LCD::Read(var, addr);
         break;
     default:
         LOG_ERROR(HW_Memory, "unknown Read%lu @ 0x%08X", sizeof(var) * 8, addr);
diff --git a/src/video_core/pica.h b/src/video_core/pica.h
index 684ec9818..9628a7589 100644
--- a/src/video_core/pica.h
+++ b/src/video_core/pica.h
@@ -372,9 +372,9 @@ struct Regs {
     INSERT_PADDING_WORDS(0x2);
 
     const std::array<Regs::TevStageConfig,6> GetTevStages() const {
-        return { tev_stage0, tev_stage1,
-                 tev_stage2, tev_stage3,
-                 tev_stage4, tev_stage5 };
+        return {{ tev_stage0, tev_stage1,
+                  tev_stage2, tev_stage3,
+                  tev_stage4, tev_stage5 }};
     };
 
     enum class BlendEquation : u32 {
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index d31c46cca..518f79331 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -94,14 +94,27 @@ void RasterizerOpenGL::InitObjects() {
     // Create textures for OGL framebuffer that will be rendered to, initially 1x1 to succeed in framebuffer creation
     fb_color_texture.texture.Create();
     ReconfigureColorTexture(fb_color_texture, Pica::Regs::ColorFormat::RGBA8, 1, 1);
+
+    state.texture_units[0].enabled_2d = true;
+    state.texture_units[0].texture_2d = fb_color_texture.texture.handle;
+    state.Apply();
+
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAX_LEVEL, 0);
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
 
+    state.texture_units[0].texture_2d = 0;
+    state.Apply();
+
     fb_depth_texture.texture.Create();
     ReconfigureDepthTexture(fb_depth_texture, Pica::Regs::DepthFormat::D16, 1, 1);
+
+    state.texture_units[0].enabled_2d = true;
+    state.texture_units[0].texture_2d = fb_depth_texture.texture.handle;
+    state.Apply();
+
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAX_LEVEL, 0);
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
@@ -110,14 +123,13 @@ void RasterizerOpenGL::InitObjects() {
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_COMPARE_FUNC, GL_LEQUAL);
     glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_COMPARE_MODE, GL_NONE);
 
+    state.texture_units[0].texture_2d = 0;
+    state.Apply();
+
     // Configure OpenGL framebuffer
     framebuffer.Create();
 
     state.draw.framebuffer = framebuffer.handle;
-
-    // Unbind texture to allow binding to framebuffer
-    state.texture_units[0].enabled_2d = true;
-    state.texture_units[0].texture_2d = 0;
     state.Apply();
 
     glActiveTexture(GL_TEXTURE0);
@@ -205,7 +217,19 @@ void RasterizerOpenGL::DrawTriangles() {
 
     vertex_batch.clear();
 
-    // TODO: Flush the resource cache at the current depth and color framebuffer addresses for render-to-texture
+    // Flush the resource cache at the current depth and color framebuffer addresses for render-to-texture
+    const auto& regs = Pica::g_state.regs;
+
+    PAddr cur_fb_color_addr = regs.framebuffer.GetColorBufferPhysicalAddress();
+    u32 cur_fb_color_size = Pica::Regs::BytesPerColorPixel(regs.framebuffer.color_format)
+                            * regs.framebuffer.GetWidth() * regs.framebuffer.GetHeight();
+
+    PAddr cur_fb_depth_addr = regs.framebuffer.GetDepthBufferPhysicalAddress();
+    u32 cur_fb_depth_size = Pica::Regs::BytesPerDepthPixel(regs.framebuffer.depth_format)
+                            * regs.framebuffer.GetWidth() * regs.framebuffer.GetHeight();
+
+    res_cache.NotifyFlush(cur_fb_color_addr, cur_fb_color_size);
+    res_cache.NotifyFlush(cur_fb_depth_addr, cur_fb_depth_size);
 }
 
 void RasterizerOpenGL::CommitFramebuffer() {
@@ -472,6 +496,9 @@ void RasterizerOpenGL::ReconfigureColorTexture(TextureInfo& texture, Pica::Regs:
     glActiveTexture(GL_TEXTURE0);
     glTexImage2D(GL_TEXTURE_2D, 0, internal_format, texture.width, texture.height, 0,
                  texture.gl_format, texture.gl_type, nullptr);
+
+    state.texture_units[0].texture_2d = 0;
+    state.Apply();
 }
 
 void RasterizerOpenGL::ReconfigureDepthTexture(DepthTextureInfo& texture, Pica::Regs::DepthFormat format, u32 width, u32 height) {
@@ -491,7 +518,7 @@ void RasterizerOpenGL::ReconfigureDepthTexture(DepthTextureInfo& texture, Pica::
     case Pica::Regs::DepthFormat::D24:
         internal_format = GL_DEPTH_COMPONENT24;
         texture.gl_format = GL_DEPTH_COMPONENT;
-        texture.gl_type = GL_UNSIGNED_INT_24_8;
+        texture.gl_type = GL_UNSIGNED_INT;
         break;
 
     case Pica::Regs::DepthFormat::D24S8:
@@ -513,6 +540,9 @@ void RasterizerOpenGL::ReconfigureDepthTexture(DepthTextureInfo& texture, Pica::
     glActiveTexture(GL_TEXTURE0);
     glTexImage2D(GL_TEXTURE_2D, 0, internal_format, texture.width, texture.height, 0,
                  texture.gl_format, texture.gl_type, nullptr);
+
+    state.texture_units[0].texture_2d = 0;
+    state.Apply();
 }
 
 void RasterizerOpenGL::SyncFramebuffer() {
@@ -652,6 +682,10 @@ void RasterizerOpenGL::SyncDepthTest() {
     const auto& regs = Pica::g_state.regs;
     state.depth.test_enabled = (regs.output_merger.depth_test_enable == 1);
     state.depth.test_func = PicaToGL::CompareFunc(regs.output_merger.depth_test_func);
+    state.color_mask.red_enabled = regs.output_merger.red_enable;
+    state.color_mask.green_enabled = regs.output_merger.green_enable;
+    state.color_mask.blue_enabled = regs.output_merger.blue_enable;
+    state.color_mask.alpha_enabled = regs.output_merger.alpha_enable;
     state.depth.write_mask = regs.output_merger.depth_write_enable ? GL_TRUE : GL_FALSE;
 }
 
@@ -759,10 +793,10 @@ void RasterizerOpenGL::ReloadColorBuffer() {
         for (int x = 0; x < fb_color_texture.width; ++x) {
             const u32 coarse_y = y & ~7;
             u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * fb_color_texture.width * bytes_per_pixel;
-            u32 gl_px_idx = x * bytes_per_pixel + y * fb_color_texture.width * bytes_per_pixel;
+            u32 gl_pixel_index = (x + y * fb_color_texture.width) * bytes_per_pixel;
 
             u8* pixel = color_buffer + dst_offset;
-            memcpy(&temp_fb_color_buffer[gl_px_idx], pixel, bytes_per_pixel);
+            memcpy(&temp_fb_color_buffer[gl_pixel_index], pixel, bytes_per_pixel);
         }
     }
 
@@ -773,6 +807,9 @@ void RasterizerOpenGL::ReloadColorBuffer() {
     glActiveTexture(GL_TEXTURE0);
     glTexSubImage2D(GL_TEXTURE_2D, 0, 0, 0, fb_color_texture.width, fb_color_texture.height,
                     fb_color_texture.gl_format, fb_color_texture.gl_type, temp_fb_color_buffer.get());
+
+    state.texture_units[0].texture_2d = 0;
+    state.Apply();
 }
 
 void RasterizerOpenGL::ReloadDepthBuffer() {
@@ -790,29 +827,29 @@ void RasterizerOpenGL::ReloadDepthBuffer() {
 
     std::unique_ptr<u8[]> temp_fb_depth_buffer(new u8[fb_depth_texture.width * fb_depth_texture.height * gl_bpp]);
 
-    for (int y = 0; y < fb_depth_texture.height; ++y) {
-        for (int x = 0; x < fb_depth_texture.width; ++x) {
-            const u32 coarse_y = y & ~7;
-            u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * fb_depth_texture.width * bytes_per_pixel;
-            u32 gl_px_idx = x + y * fb_depth_texture.width;
-
-            switch (fb_depth_texture.format) {
-            case Pica::Regs::DepthFormat::D16:
-                ((u16*)temp_fb_depth_buffer.get())[gl_px_idx] = Color::DecodeD16(depth_buffer + dst_offset);
-                break;
-            case Pica::Regs::DepthFormat::D24:
-                ((u32*)temp_fb_depth_buffer.get())[gl_px_idx] = Color::DecodeD24(depth_buffer + dst_offset);
-                break;
-            case Pica::Regs::DepthFormat::D24S8:
-            {
-                Math::Vec2<u32> depth_stencil = Color::DecodeD24S8(depth_buffer + dst_offset);
-                ((u32*)temp_fb_depth_buffer.get())[gl_px_idx] = (depth_stencil.x << 8) | depth_stencil.y;
-                break;
+    u8* temp_fb_depth_data = bytes_per_pixel == 3 ? (temp_fb_depth_buffer.get() + 1) : temp_fb_depth_buffer.get();
+
+    if (fb_depth_texture.format == Pica::Regs::DepthFormat::D24S8) {
+        for (int y = 0; y < fb_depth_texture.height; ++y) {
+            for (int x = 0; x < fb_depth_texture.width; ++x) {
+                const u32 coarse_y = y & ~7;
+                u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * fb_depth_texture.width * bytes_per_pixel;
+                u32 gl_pixel_index = (x + y * fb_depth_texture.width);
+
+                u8* pixel = depth_buffer + dst_offset;
+                u32 depth_stencil = *(u32*)pixel;
+                ((u32*)temp_fb_depth_data)[gl_pixel_index] = (depth_stencil << 8) | (depth_stencil >> 24);
             }
-            default:
-                LOG_CRITICAL(Render_OpenGL, "Unknown memory framebuffer depth format %x", fb_depth_texture.format);
-                UNIMPLEMENTED();
-                break;
+        }
+    } else {
+        for (int y = 0; y < fb_depth_texture.height; ++y) {
+            for (int x = 0; x < fb_depth_texture.width; ++x) {
+                const u32 coarse_y = y & ~7;
+                u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * fb_depth_texture.width * bytes_per_pixel;
+                u32 gl_pixel_index = (x + y * fb_depth_texture.width) * gl_bpp;
+
+                u8* pixel = depth_buffer + dst_offset;
+                memcpy(&temp_fb_depth_data[gl_pixel_index], pixel, bytes_per_pixel);
             }
         }
     }
@@ -824,6 +861,9 @@ void RasterizerOpenGL::ReloadDepthBuffer() {
     glActiveTexture(GL_TEXTURE0);
     glTexSubImage2D(GL_TEXTURE_2D, 0, 0, 0, fb_depth_texture.width, fb_depth_texture.height,
                     fb_depth_texture.gl_format, fb_depth_texture.gl_type, temp_fb_depth_buffer.get());
+
+    state.texture_units[0].texture_2d = 0;
+    state.Apply();
 }
 
 void RasterizerOpenGL::CommitColorBuffer() {
@@ -842,15 +882,18 @@ void RasterizerOpenGL::CommitColorBuffer() {
             glActiveTexture(GL_TEXTURE0);
             glGetTexImage(GL_TEXTURE_2D, 0, fb_color_texture.gl_format, fb_color_texture.gl_type, temp_gl_color_buffer.get());
 
+            state.texture_units[0].texture_2d = 0;
+            state.Apply();
+
             // Directly copy pixels. Internal OpenGL color formats are consistent so no conversion is necessary.
             for (int y = 0; y < fb_color_texture.height; ++y) {
                 for (int x = 0; x < fb_color_texture.width; ++x) {
                     const u32 coarse_y = y & ~7;
                     u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * fb_color_texture.width * bytes_per_pixel;
-                    u32 gl_px_idx = x * bytes_per_pixel + y * fb_color_texture.width * bytes_per_pixel;
+                    u32 gl_pixel_index = x * bytes_per_pixel + y * fb_color_texture.width * bytes_per_pixel;
 
                     u8* pixel = color_buffer + dst_offset;
-                    memcpy(pixel, &temp_gl_color_buffer[gl_px_idx], bytes_per_pixel);
+                    memcpy(pixel, &temp_gl_color_buffer[gl_pixel_index], bytes_per_pixel);
                 }
             }
         }
@@ -877,29 +920,32 @@ void RasterizerOpenGL::CommitDepthBuffer() {
             glActiveTexture(GL_TEXTURE0);
             glGetTexImage(GL_TEXTURE_2D, 0, fb_depth_texture.gl_format, fb_depth_texture.gl_type, temp_gl_depth_buffer.get());
 
-            for (int y = 0; y < fb_depth_texture.height; ++y) {
-                for (int x = 0; x < fb_depth_texture.width; ++x) {
-                    const u32 coarse_y = y & ~7;
-                    u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * fb_depth_texture.width * bytes_per_pixel;
-                    u32 gl_px_idx = x + y * fb_depth_texture.width;
-
-                    switch (fb_depth_texture.format) {
-                    case Pica::Regs::DepthFormat::D16:
-                        Color::EncodeD16(((u16*)temp_gl_depth_buffer.get())[gl_px_idx], depth_buffer + dst_offset);
-                        break;
-                    case Pica::Regs::DepthFormat::D24:
-                        Color::EncodeD24(((u32*)temp_gl_depth_buffer.get())[gl_px_idx], depth_buffer + dst_offset);
-                        break;
-                    case Pica::Regs::DepthFormat::D24S8:
-                    {
-                        u32 depth_stencil = ((u32*)temp_gl_depth_buffer.get())[gl_px_idx];
-                        Color::EncodeD24S8((depth_stencil >> 8), depth_stencil & 0xFF, depth_buffer + dst_offset);
-                        break;
+            state.texture_units[0].texture_2d = 0;
+            state.Apply();
+
+            u8* temp_gl_depth_data = bytes_per_pixel == 3 ? (temp_gl_depth_buffer.get() + 1) : temp_gl_depth_buffer.get();
+
+            if (fb_depth_texture.format == Pica::Regs::DepthFormat::D24S8) {
+                for (int y = 0; y < fb_depth_texture.height; ++y) {
+                    for (int x = 0; x < fb_depth_texture.width; ++x) {
+                        const u32 coarse_y = y & ~7;
+                        u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * fb_depth_texture.width * bytes_per_pixel;
+                        u32 gl_pixel_index = (x + y * fb_depth_texture.width);
+
+                        u8* pixel = depth_buffer + dst_offset;
+                        u32 depth_stencil = ((u32*)temp_gl_depth_data)[gl_pixel_index];
+                        *(u32*)pixel = (depth_stencil >> 8) | (depth_stencil << 24);
                     }
-                    default:
-                        LOG_CRITICAL(Render_OpenGL, "Unknown framebuffer depth format %x", fb_depth_texture.format);
-                        UNIMPLEMENTED();
-                        break;
+                }
+            } else {
+                for (int y = 0; y < fb_depth_texture.height; ++y) {
+                    for (int x = 0; x < fb_depth_texture.width; ++x) {
+                        const u32 coarse_y = y & ~7;
+                        u32 dst_offset = VideoCore::GetMortonOffset(x, y, bytes_per_pixel) + coarse_y * fb_depth_texture.width * bytes_per_pixel;
+                        u32 gl_pixel_index = (x + y * fb_depth_texture.width) * gl_bpp;
+
+                        u8* pixel = depth_buffer + dst_offset;
+                        memcpy(pixel, &temp_gl_depth_data[gl_pixel_index], bytes_per_pixel);
                     }
                 }
             }
diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp
index 9c5f38f94..3526e16d5 100644
--- a/src/video_core/renderer_opengl/gl_state.cpp
+++ b/src/video_core/renderer_opengl/gl_state.cpp
@@ -16,6 +16,11 @@ OpenGLState::OpenGLState() {
     depth.test_func = GL_LESS;
     depth.write_mask = GL_TRUE;
 
+    color_mask.red_enabled = GL_TRUE;
+    color_mask.green_enabled = GL_TRUE;
+    color_mask.blue_enabled = GL_TRUE;
+    color_mask.alpha_enabled = GL_TRUE;
+
     stencil.test_enabled = false;
     stencil.test_func = GL_ALWAYS;
     stencil.test_ref = 0;
@@ -77,6 +82,15 @@ void OpenGLState::Apply() {
         glDepthMask(depth.write_mask);
     }
 
+    // Color mask
+    if (color_mask.red_enabled != cur_state.color_mask.red_enabled ||
+            color_mask.green_enabled != cur_state.color_mask.green_enabled ||
+            color_mask.blue_enabled != cur_state.color_mask.blue_enabled ||
+            color_mask.alpha_enabled != cur_state.color_mask.alpha_enabled) {
+        glColorMask(color_mask.red_enabled, color_mask.green_enabled,
+                    color_mask.blue_enabled, color_mask.alpha_enabled);
+    }
+
     // Stencil test
     if (stencil.test_enabled != cur_state.stencil.test_enabled) {
         if (stencil.test_enabled) {
@@ -87,8 +101,8 @@ void OpenGLState::Apply() {
     }
 
     if (stencil.test_func != cur_state.stencil.test_func ||
-        stencil.test_ref != cur_state.stencil.test_ref ||
-        stencil.test_mask != cur_state.stencil.test_mask) {
+            stencil.test_ref != cur_state.stencil.test_ref ||
+            stencil.test_mask != cur_state.stencil.test_mask) {
         glStencilFunc(stencil.test_func, stencil.test_ref, stencil.test_mask);
     }
 
@@ -112,17 +126,19 @@ void OpenGLState::Apply() {
     }
 
     if (blend.color.red != cur_state.blend.color.red ||
-        blend.color.green != cur_state.blend.color.green ||
-        blend.color.blue != cur_state.blend.color.blue ||
-        blend.color.alpha != cur_state.blend.color.alpha) {
-        glBlendColor(blend.color.red, blend.color.green, blend.color.blue, blend.color.alpha);
+            blend.color.green != cur_state.blend.color.green ||
+            blend.color.blue != cur_state.blend.color.blue ||
+            blend.color.alpha != cur_state.blend.color.alpha) {
+        glBlendColor(blend.color.red, blend.color.green,
+                     blend.color.blue, blend.color.alpha);
     }
 
     if (blend.src_rgb_func != cur_state.blend.src_rgb_func ||
-        blend.dst_rgb_func != cur_state.blend.dst_rgb_func ||
-        blend.src_a_func != cur_state.blend.src_a_func ||
-        blend.dst_a_func != cur_state.blend.dst_a_func) {
-        glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func, blend.src_a_func, blend.dst_a_func);
+            blend.dst_rgb_func != cur_state.blend.dst_rgb_func ||
+            blend.src_a_func != cur_state.blend.src_a_func ||
+            blend.dst_a_func != cur_state.blend.dst_a_func) {
+        glBlendFuncSeparate(blend.src_rgb_func, blend.dst_rgb_func,
+                            blend.src_a_func, blend.dst_a_func);
     }
 
     if (logic_op != cur_state.logic_op) {
diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h
index 6b97721d6..26b916360 100644
--- a/src/video_core/renderer_opengl/gl_state.h
+++ b/src/video_core/renderer_opengl/gl_state.h
@@ -20,6 +20,13 @@ public:
     } depth;
 
     struct {
+        GLboolean red_enabled;
+        GLboolean green_enabled;
+        GLboolean blue_enabled;
+        GLboolean alpha_enabled;
+    } color_mask; // GL_COLOR_WRITEMASK
+
+    struct {
         bool test_enabled; // GL_STENCIL_TEST
         GLenum test_func; // GL_STENCIL_FUNC
         GLint test_ref; // GL_STENCIL_REF
diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp
index 382aeaa05..3399ca123 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.cpp
+++ b/src/video_core/renderer_opengl/renderer_opengl.cpp
@@ -170,6 +170,9 @@ void RendererOpenGL::LoadFBToActiveGLTexture(const GPU::Regs::FramebufferConfig&
                     texture.gl_format, texture.gl_type, framebuffer_data);
 
     glPixelStorei(GL_UNPACK_ROW_LENGTH, 0);
+
+    state.texture_units[0].texture_2d = 0;
+    state.Apply();
 }
 
 /**
@@ -239,6 +242,9 @@ void RendererOpenGL::InitOpenGLObjects() {
         glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
     }
 
+    state.texture_units[0].texture_2d = 0;
+    state.Apply();
+
     hw_rasterizer->InitObjects();
 }
 
@@ -370,6 +376,8 @@ void RendererOpenGL::Init() {
     }
 
     LOG_INFO(Render_OpenGL, "GL_VERSION: %s", glGetString(GL_VERSION));
+    LOG_INFO(Render_OpenGL, "GL_VENDOR: %s", glGetString(GL_VENDOR));
+    LOG_INFO(Render_OpenGL, "GL_RENDERER: %s", glGetString(GL_RENDERER));
     InitOpenGLObjects();
 }