diff --git a/Core/HLE/HLE.cpp b/Core/HLE/HLE.cpp
index faeea8ead4..4930f92723 100644
--- a/Core/HLE/HLE.cpp
+++ b/Core/HLE/HLE.cpp
@@ -21,6 +21,7 @@
 #include <vector>
 #include "../MemMap.h"
 #include "../Config.h"
+#include "Core/CoreTiming.h"
 #include "Core/Reporting.h"
 
 #include "HLETables.h"
@@ -55,12 +56,27 @@ enum
 static std::vector<HLEModule> moduleDB;
 static std::vector<Syscall> unresolvedSyscalls;
 static std::vector<Syscall> exportedCalls;
+static int delayedResultEvent = -1;
 static int hleAfterSyscall = HLE_AFTER_NOTHING;
 static const char *hleAfterSyscallReschedReason;
 
+void hleDelayResultFinish(u64 userdata, int cycleslate)
+{
+	u32 error;
+	SceUID threadID = (SceUID) userdata;
+	SceUID verify = __KernelGetWaitID(threadID, WAITTYPE_DELAY, error);
+	SceUID result = __KernelGetWaitValue(threadID, error);
+
+	if (error == 0 && verify == 1)
+		__KernelResumeThreadFromWait(threadID, result);
+	else
+		WARN_LOG(HLE, "Someone else woke up HLE-blocked thread?");
+}
+
 void HLEInit()
 {
 	RegisterAllModules();
+	delayedResultEvent = CoreTiming::RegisterEvent("HLEDelayedResult", hleDelayResultFinish);
 }
 
 void HLEDoState(PointerWrap &p)
@@ -68,6 +84,8 @@ void HLEDoState(PointerWrap &p)
 	Syscall sc = {""};
 	p.Do(unresolvedSyscalls, sc);
 	p.Do(exportedCalls, sc);
+	p.Do(delayedResultEvent);
+	CoreTiming::RestoreRegisterEvent(delayedResultEvent, "HLEDelayedResult", hleDelayResultFinish);
 	p.DoMarker("HLE");
 }
 
@@ -310,6 +328,19 @@ bool hleExecuteDebugBreak(const HLEFunction &func)
 	return true;
 }
 
+u32 hleDelayResult(u32 result, const char *reason, int usec)
+{
+	CoreTiming::ScheduleEvent(usToCycles(usec), delayedResultEvent, __KernelGetCurThread());
+	__KernelWaitCurThread(WAITTYPE_DELAY, 1, result, 0, false, reason);
+	return result;
+}
+
+void hleEatMicro(int usec)
+{
+	// Maybe this should Idle, at least for larger delays?  Could that cause issues?
+	currentMIPS->downcount -= (int) usToCycles(usec);
+}
+
 inline void hleFinishSyscall(int modulenum, int funcnum)
 {
 	if ((hleAfterSyscall & HLE_AFTER_CURRENT_CALLBACKS) != 0)
diff --git a/Core/HLE/HLE.h b/Core/HLE/HLE.h
index 40b4a17dbc..2618eefcfb 100644
--- a/Core/HLE/HLE.h
+++ b/Core/HLE/HLE.h
@@ -87,6 +87,10 @@ void hleRunInterrupts();
 // Pause emulation after the syscall finishes.
 void hleDebugBreak();
 
+// Delays the result for usec microseconds, allowing other threads to run during this time.
+u32 hleDelayResult(u32 result, const char *reason, int usec);
+void hleEatMicro(int usec);
+
 void HLEInit();
 void HLEDoState(PointerWrap &p);
 void HLEShutdown();
diff --git a/Core/HLE/sceDisplay.cpp b/Core/HLE/sceDisplay.cpp
index 651a5adbd1..2eebe46f1a 100644
--- a/Core/HLE/sceDisplay.cpp
+++ b/Core/HLE/sceDisplay.cpp
@@ -543,6 +543,7 @@ u32 sceDisplayWaitVblank() {
 		return 0;
 	} else {
 		DEBUG_LOG(HLE,"sceDisplayWaitVblank() - not waiting since in vBlank");
+		hleEatMicro(5);
 		return 1;
 	}
 }
@@ -562,6 +563,7 @@ u32 sceDisplayWaitVblankCB() {
 		return 0;
 	} else {
 		DEBUG_LOG(HLE,"sceDisplayWaitVblank() - not waiting since in vBlank");
+		hleEatMicro(5);
 		return 1;
 	}
 }
@@ -584,12 +586,7 @@ u32 sceDisplayGetVcount() {
 	// Too spammy
 	// DEBUG_LOG(HLE,"%i=sceDisplayGetVcount()", vCount);
 
-	// Puyo Puyo Fever polls this as a substitute for waiting for vblank.
-	// As a result, the game never gets to reschedule so it doesn't mix audio and things break.
-	// Need to find a better hack as this breaks games like Project Diva.
-	// hleReSchedule("sceDisplayGetVcount hack");  // Puyo puyo hack?
-
-	CoreTiming::Idle(1000000);
+	hleEatMicro(2);
 	return vCount;
 }
 
diff --git a/Core/HLE/sceKernelModule.cpp b/Core/HLE/sceKernelModule.cpp
index f60ff6e215..bb23a869df 100644
--- a/Core/HLE/sceKernelModule.cpp
+++ b/Core/HLE/sceKernelModule.cpp
@@ -230,24 +230,13 @@ struct SceKernelSMOption {
 //////////////////////////////////////////////////////////////////////////
 // STATE BEGIN
 static int actionAfterModule;
-static int eventLoadModule = -1;
 static SceUID mainModuleID;	// hack
 // STATE END
 //////////////////////////////////////////////////////////////////////////
 
-void __KernelModuleLoaded(u64 userdata, int cycleslate)
-{
-	u32 error;
-	SceUID threadID = userdata & 0xFFFFFFFF;
-	SceUID moduleID = __KernelGetWaitValue(threadID, error);
-	if (error == 0 && moduleID != 0)
-		__KernelResumeThreadFromWait(threadID, moduleID);
-}
-
 void __KernelModuleInit()
 {
 	actionAfterModule = __KernelRegisterActionType(AfterModuleEntryCall::Create);
-	eventLoadModule = CoreTiming::RegisterEvent("LoadModule", __KernelModuleLoaded);
 }
 
 void __KernelModuleDoState(PointerWrap &p)
@@ -255,8 +244,6 @@ void __KernelModuleDoState(PointerWrap &p)
 	p.Do(mainModuleID);
 	p.Do(actionAfterModule);
 	__KernelRestoreActionType(actionAfterModule, AfterModuleEntryCall::Create);
-	p.Do(eventLoadModule);
-	CoreTiming::RestoreRegisterEvent(eventLoadModule, "LoadModule", __KernelModuleLoaded);
 	p.DoMarker("sceKernelModule");
 }
 
@@ -827,9 +814,7 @@ u32 sceKernelLoadModule(const char *name, u32 flags, u32 optionAddr)
 	}
 
 	// TODO: This is not the right timing and probably not the right wait type, just an approximation.
-	CoreTiming::ScheduleEvent(usToCycles(500), eventLoadModule, __KernelGetCurThread());
-	__KernelWaitCurThread(WAITTYPE_SEMA, -1, module->GetUID(), 0, false, "module loaded");
-	return module->GetUID();
+	return hleDelayResult(module->GetUID(), "module loaded", 500);
 }
 
 void sceKernelStartModule(u32 moduleId, u32 argsize, u32 argAddr, u32 returnValueAddr, u32 optionAddr)
diff --git a/Core/HLE/sceUtility.cpp b/Core/HLE/sceUtility.cpp
index e3c15bca52..677b271fac 100644
--- a/Core/HLE/sceUtility.cpp
+++ b/Core/HLE/sceUtility.cpp
@@ -17,6 +17,7 @@
 
 #include "HLE.h"
 #include "../MIPS/MIPS.h"
+#include "Core/Reporting.h"
 
 #include "sceKernel.h"
 #include "sceKernelThread.h"
@@ -30,6 +31,7 @@
 #include "../Dialog/PSPOskDialog.h"
 
 const int SCE_ERROR_MODULE_BAD_ID = 0x80111101;
+const int SCE_ERROR_AV_MODULE_BAD_ID = 0x80110F01;
 
 PSPSaveDialog saveDialog;
 PSPMsgDialog msgDialog;
@@ -92,16 +94,21 @@ int sceUtilitySavedataUpdate(int animSpeed)
 
 u32 sceUtilityLoadAvModule(u32 module)
 {
+	if (module > 7)
+	{
+		ERROR_LOG(HLE, "sceUtilityLoadAvModule(%i): invalid module id", module);
+		Reporting::ReportMessage("sceUtilityLoadAvModule(%i): invalid module id", module);
+		return SCE_ERROR_AV_MODULE_BAD_ID;
+	}
+
 	DEBUG_LOG(HLE,"sceUtilityLoadAvModule(%i)", module);
-	hleReSchedule("utilityloadavmodule");
-	return 0;
+	return hleDelayResult(0, "utility av module loaded", 25000);
 }
 
 u32 sceUtilityUnloadAvModule(u32 module)
 {
 	DEBUG_LOG(HLE,"sceUtilityUnloadAvModule(%i)", module);
-	hleReSchedule("utilityunloadavmodule");
-	return 0;
+	return hleDelayResult(0, "utility av module unloaded", 800);
 }
 
 u32 sceUtilityLoadModule(u32 module)
@@ -110,6 +117,7 @@ u32 sceUtilityLoadModule(u32 module)
 	if (module < 0x100 || module > 0x601)
 	{
 		ERROR_LOG(HLE, "sceUtilityLoadModule(%i): invalid module id", module);
+		Reporting::ReportMessage("sceUtilityLoadModule(%i): invalid module id", module);
 		return SCE_ERROR_MODULE_BAD_ID;
 	}
 
@@ -117,10 +125,9 @@ u32 sceUtilityLoadModule(u32 module)
 	// TODO: Each module has its own timing, technically, but this is a low-end.
 	// Note: Some modules have dependencies, but they still resched.
 	if (module == 0x3FF)
-		sceKernelDelayThread(130);
+		return hleDelayResult(0, "utility module loaded", 130);
 	else
-		sceKernelDelayThread(25000);
-	return 0;
+		return hleDelayResult(0, "utility module loaded", 25000);
 }
 
 u32 sceUtilityUnloadModule(u32 module)
@@ -129,6 +136,7 @@ u32 sceUtilityUnloadModule(u32 module)
 	if (module < 0x100 || module > 0x601)
 	{
 		ERROR_LOG(HLE, "sceUtilityUnloadModule(%i): invalid module id", module);
+		Reporting::ReportMessage("sceUtilityUnloadModule(%i): invalid module id", module);
 		return SCE_ERROR_MODULE_BAD_ID;
 	}
 
@@ -136,10 +144,9 @@ u32 sceUtilityUnloadModule(u32 module)
 	// TODO: Each module has its own timing, technically, but this is a low-end.
 	// Note: If not loaded, it should not reschedule actually...
 	if (module == 0x3FF)
-		sceKernelDelayThread(110);
+		return hleDelayResult(0, "utility module unloaded", 110);
 	else
-		sceKernelDelayThread(400);
-	return 0;
+		return hleDelayResult(0, "utility module unloaded", 400);
 }
 
 int sceUtilityMsgDialogInitStart(u32 structAddr)
diff --git a/android/jni/TestRunner.cpp b/android/jni/TestRunner.cpp
index 6170711f2c..3783081a1f 100644
--- a/android/jni/TestRunner.cpp
+++ b/android/jni/TestRunner.cpp
@@ -71,6 +71,10 @@ void RunTests()
 	coreParam.useMediaEngine = false;
 	coreParam.collectEmuLog = &output;
 
+	// Never report from tests.
+	std::string savedReportHost = g_Config.sReportHost;
+	g_Config.sReportHost = "";
+
 	for (int i = 0; i < ARRAY_SIZE(testsToRun); i++) {
 		const char *testName = testsToRun[i];
 		coreParam.fileToStart = g_Config.memCardDirectory + "pspautotests/tests/" + testName + ".prx";
@@ -136,4 +140,6 @@ void RunTests()
 	}
 	glstate.Restore();
 	glstate.viewport.set(0,0,pixel_xres,pixel_yres);
+
+	g_Config.sReportHost = savedReportHost;
 }
\ No newline at end of file
diff --git a/headless/Headless.cpp b/headless/Headless.cpp
index f8db1ff0fb..96ddf00612 100644
--- a/headless/Headless.cpp
+++ b/headless/Headless.cpp
@@ -166,6 +166,8 @@ int main(int argc, const char* argv[])
 	g_Config.bEnableSound = false;
 	g_Config.bFirstRun = false;
 	g_Config.bIgnoreBadMemAccess = true;
+	// Never report from tests.
+	g_Config.sReportHost = "";
 
 #if defined(ANDROID)
 #elif defined(BLACKBERRY) || defined(__SYMBIAN32__)
diff --git a/test.py b/test.py
index d2836664b5..625b1d5c93 100755
--- a/test.py
+++ b/test.py
@@ -53,6 +53,7 @@ tests_good = [
   "ctrl/sampling/sampling",
   "ctrl/sampling2/sampling2",
   "display/display",
+  "display/vblankmulti",
   "dmac/dmactest",
   "loader/bss/bss",
   "intr/intr",
@@ -134,7 +135,6 @@ tests_next = [
   "audio/atrac/atractest",
   "audio/mp3/mp3test",
   "audio/sascore/sascore",
-  "display/vblankmulti",
   "malloc/malloc",
   "threads/fpl/fpl",
   "threads/k0/k0",