23 files changed, 486 insertions, 177 deletions

diff --git a/indra/llcommon/CMakeLists.txt b/indra/llcommon/CMakeLists.txt
index 72a997cd89..66995d8083 100755
--- a/indra/llcommon/CMakeLists.txt
+++ b/indra/llcommon/CMakeLists.txt
@@ -122,6 +122,7 @@ set(llcommon_HEADER_FILES
     linden_common.h
     linked_lists.h
     llaccountingcost.h
+    llalignedarray.h
     llallocator.h
     llallocator_heap_profile.h
     llagentconstants.h
@@ -236,6 +237,7 @@ set(llcommon_HEADER_FILES
     llstrider.h
     llstring.h
     llstringtable.h
+    llstaticstringtable.h
     llsys.h
     llthread.h
     llthreadsafequeue.h
diff --git a/indra/llcommon/llalignedarray.h b/indra/llcommon/llalignedarray.h
new file mode 100644
index 0000000000..ed8fd31205
--- /dev/null
+++ b/indra/llcommon/llalignedarray.h
@@ -0,0 +1,139 @@
+/** 
+ * @file llalignedarray.h
+ * @brief A static array which obeys alignment restrictions and mimics std::vector accessors.
+ *
+ * $LicenseInfo:firstyear=2002&license=viewerlgpl$
+ * Second Life Viewer Source Code
+ * Copyright (C) 2010, Linden Research, Inc.
+ * 
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation;
+ * version 2.1 of the License only.
+ * 
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ * 
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
+ * 
+ * Linden Research, Inc., 945 Battery Street, San Francisco, CA  94111  USA
+ * $/LicenseInfo$
+ */
+
+#ifndef LL_LLALIGNEDARRAY_H
+#define LL_LLALIGNEDARRAY_H
+
+#include "llmemory.h"
+
+template <class T, U32 alignment>
+class LLAlignedArray
+{
+public:
+	T* mArray;
+	U32 mElementCount;
+	U32 mCapacity;
+
+	LLAlignedArray();
+	~LLAlignedArray();
+
+	void push_back(const T& elem);
+	U32 size() const { return mElementCount; }
+	void resize(U32 size);
+	T* append(S32 N);
+	T& operator[](int idx);
+	const T& operator[](int idx) const;
+};
+
+template <class T, U32 alignment>
+LLAlignedArray<T, alignment>::LLAlignedArray()
+{
+	llassert(alignment >= 16);
+	mArray = NULL;
+	mElementCount = 0;
+	mCapacity = 0;
+}
+
+template <class T, U32 alignment>
+LLAlignedArray<T, alignment>::~LLAlignedArray()
+{
+	ll_aligned_free(mArray);
+	mArray = NULL;
+	mElementCount = 0;
+	mCapacity = 0;
+}
+
+template <class T, U32 alignment>
+void LLAlignedArray<T, alignment>::push_back(const T& elem)
+{
+	T* old_buf = NULL;
+	if (mCapacity <= mElementCount)
+	{
+		mCapacity++;
+		mCapacity *= 2;
+		T* new_buf = (T*) ll_aligned_malloc(mCapacity*sizeof(T), alignment);
+		if (mArray)
+		{
+			ll_memcpy_nonaliased_aligned_16((char*)new_buf, (char*)mArray, sizeof(T)*mElementCount);
+		}
+		old_buf = mArray;
+		mArray = new_buf;
+	}
+
+	mArray[mElementCount++] = elem;
+
+	//delete old array here to prevent error on a.push_back(a[0])
+	ll_aligned_free(old_buf);
+}
+
+template <class T, U32 alignment>
+void LLAlignedArray<T, alignment>::resize(U32 size)
+{
+	if (mCapacity < size)
+	{
+		mCapacity = size+mCapacity*2;
+		T* new_buf = mCapacity > 0 ? (T*) ll_aligned_malloc(mCapacity*sizeof(T), alignment) : NULL;
+		if (mArray)
+		{
+			ll_memcpy_nonaliased_aligned_16((char*) new_buf, (char*) mArray, sizeof(T)*mElementCount);
+			ll_aligned_free(mArray);
+		}
+
+		/*for (U32 i = mElementCount; i < mCapacity; ++i)
+		{
+			new(new_buf+i) T();
+		}*/
+		mArray = new_buf;
+	}
+
+	mElementCount = size;
+}
+
+
+template <class T, U32 alignment>
+T& LLAlignedArray<T, alignment>::operator[](int idx)
+{
+	llassert(idx < mElementCount);
+	return mArray[idx];
+}
+
+template <class T, U32 alignment>
+const T& LLAlignedArray<T, alignment>::operator[](int idx) const
+{
+	llassert(idx < mElementCount);
+	return mArray[idx];
+}
+
+template <class T, U32 alignment>
+T* LLAlignedArray<T, alignment>::append(S32 N)
+{
+	U32 sz = size();
+	resize(sz+N);
+	return &((*this)[sz]);
+}
+
+#endif
+
diff --git a/indra/llcommon/llapr.cpp b/indra/llcommon/llapr.cpp
index a0802c6adf..b7815b0e35 100755
--- a/indra/llcommon/llapr.cpp
+++ b/indra/llcommon/llapr.cpp
@@ -226,7 +226,7 @@ void LLVolatileAPRPool::clearVolatileAPRPool()
 		llassert_always(mNumActiveRef > 0) ;
 	}
 
-	llassert(mNumTotalRef < (FULL_VOLATILE_APR_POOL << 2)) ;
+	llassert(mNumTotalRef <= (FULL_VOLATILE_APR_POOL << 2)) ;
 }
 
 BOOL LLVolatileAPRPool::isFull()
diff --git a/indra/llcommon/llapr.h b/indra/llcommon/llapr.h
index 752574c65d..1fe7d98472 100755
--- a/indra/llcommon/llapr.h
+++ b/indra/llcommon/llapr.h
@@ -164,14 +164,20 @@ public:
 	~LLAtomic32<Type>() {};
 
 	operator const Type() { apr_uint32_t data = apr_atomic_read32(&mData); return Type(data); }
+	
+	Type	CurrentValue() const { apr_uint32_t data = apr_atomic_read32(const_cast< volatile apr_uint32_t* >(&mData)); return Type(data); }
+
 	Type operator =(const Type& x) { apr_atomic_set32(&mData, apr_uint32_t(x)); return Type(mData); }
 	void operator -=(Type x) { apr_atomic_sub32(&mData, apr_uint32_t(x)); }
 	void operator +=(Type x) { apr_atomic_add32(&mData, apr_uint32_t(x)); }
 	Type operator ++(int) { return apr_atomic_inc32(&mData); } // Type++
 	Type operator --(int) { return apr_atomic_dec32(&mData); } // approximately --Type (0 if final is 0, non-zero otherwise)
+
+	Type operator ++() { return apr_atomic_inc32(&mData); } // Type++
+	Type operator --() { return apr_atomic_dec32(&mData); } // approximately --Type (0 if final is 0, non-zero otherwise)
 	
 private:
-	apr_uint32_t mData;
+	volatile apr_uint32_t mData;
 };
 
 typedef LLAtomic32<U32> LLAtomicU32;
diff --git a/indra/llcommon/llcommon.cpp b/indra/llcommon/llcommon.cpp
index 8be9e4f4de..b938b0e65a 100755
--- a/indra/llcommon/llcommon.cpp
+++ b/indra/llcommon/llcommon.cpp
@@ -44,6 +44,7 @@ void LLCommon::initClass()
 	}
 	LLTimer::initClass();
 	LLThreadSafeRefCount::initThreadSafeRefCount();
+	assert_main_thread();		// Make sure we record the main thread
 // 	LLWorkerThread::initClass();
 // 	LLFrameCallbackManager::initClass();
 }
diff --git a/indra/llcommon/llcriticaldamp.cpp b/indra/llcommon/llcriticaldamp.cpp
index 87d79b1ee0..49aac9ce75 100755
--- a/indra/llcommon/llcriticaldamp.cpp
+++ b/indra/llcommon/llcriticaldamp.cpp
@@ -87,3 +87,4 @@ F32 LLCriticalDamp::getInterpolant(const F32 time_constant, BOOL use_cache)
 
 	return interpolant;
 }
+
diff --git a/indra/llcommon/lleventapi.h b/indra/llcommon/lleventapi.h
index 1a37d780b6..5991fe8fd5 100755
--- a/indra/llcommon/lleventapi.h
+++ b/indra/llcommon/lleventapi.h
@@ -47,6 +47,7 @@ class LL_COMMON_API LLEventAPI: public LLDispatchListener,
     typedef LLInstanceTracker<LLEventAPI, std::string> ibase;
 
 public:
+
     /**
      * @param name LLEventPump name on which this LLEventAPI will listen. This
      * also serves as the LLInstanceTracker instance key.
diff --git a/indra/llcommon/lleventtimer.h b/indra/llcommon/lleventtimer.h
index 7f42623d01..dc918121e1 100755
--- a/indra/llcommon/lleventtimer.h
+++ b/indra/llcommon/lleventtimer.h
@@ -36,6 +36,7 @@
 class LL_COMMON_API LLEventTimer : public LLInstanceTracker<LLEventTimer>
 {
 public:
+
 	LLEventTimer(F32 period);	// period is the amount of time between each call to tick() in seconds
 	LLEventTimer(const LLDate& time);
 	virtual ~LLEventTimer();
diff --git a/indra/llcommon/llfasttimer.cpp b/indra/llcommon/llfasttimer.cpp
index 024fdd1b4d..01b6e60d2b 100755
--- a/indra/llcommon/llfasttimer.cpp
+++ b/indra/llcommon/llfasttimer.cpp
@@ -107,17 +107,13 @@ class NamedTimerFactory : public LLSingleton<NamedTimerFactory>
 {
 public:
 	NamedTimerFactory()
-	:	mTimerRoot(NULL)
-	{}
-
-	/*virtual */ void initSingleton()
+	:	mTimerRoot(new LLFastTimer::NamedTimer("root"))
 	{
-		mTimerRoot = new LLFastTimer::NamedTimer("root");
 		mRootFrameState.setNamedTimer(mTimerRoot);
 		mTimerRoot->setFrameState(&mRootFrameState);
 		mTimerRoot->mParent = mTimerRoot;
 		mTimerRoot->setCollapsed(false);
-		mRootFrameState.mParent = &mRootFrameState;
+		mRootFrameState.mParent = &mRootFrameState;	
 	}
 
 	~NamedTimerFactory()
diff --git a/indra/llcommon/llfasttimer.h b/indra/llcommon/llfasttimer.h
index 81c4b78775..a99a1d88af 100755
--- a/indra/llcommon/llfasttimer.h
+++ b/indra/llcommon/llfasttimer.h
@@ -139,10 +139,11 @@ public:
 
 	// used to statically declare a new named timer
 	class LL_COMMON_API DeclareTimer
-	:	public LLInstanceTracker<DeclareTimer>
+	:	public LLInstanceTracker< DeclareTimer >
 	{
 		friend class LLFastTimer;
 	public:
+
 		DeclareTimer(const std::string& name, bool open);
 		DeclareTimer(const std::string& name);
 
diff --git a/indra/llcommon/llfoldertype.h b/indra/llcommon/llfoldertype.h
index a0c847914f..a0c847914f 100755..100644
--- a/indra/llcommon/llfoldertype.h
+++ b/indra/llcommon/llfoldertype.h
diff --git a/indra/llcommon/llinitparam.h b/indra/llcommon/llinitparam.h
index ae836645b9..03ab0fb67f 100755
--- a/indra/llcommon/llinitparam.h
+++ b/indra/llcommon/llinitparam.h
@@ -1952,7 +1952,7 @@ namespace LLInitParam
 		class Mandatory : public TypedParam<T, NAME_VALUE_LOOKUP, false>
 		{
 			typedef TypedParam<T, NAME_VALUE_LOOKUP, false>		super_t;
-			typedef Mandatory<T, NAME_VALUE_LOOKUP>										self_t;
+			typedef Mandatory<T, NAME_VALUE_LOOKUP>				self_t;
 			typedef typename super_t::value_t					value_t;
 			typedef typename super_t::default_value_t			default_value_t;
 
diff --git a/indra/llcommon/llinstancetracker.cpp b/indra/llcommon/llinstancetracker.cpp
index 5dc3ea5d7b..64a313b5ff 100755
--- a/indra/llcommon/llinstancetracker.cpp
+++ b/indra/llcommon/llinstancetracker.cpp
@@ -32,18 +32,3 @@
 // external library headers
 // other Linden headers
 
-//static 
-void * & LLInstanceTrackerBase::getInstances(std::type_info const & info)
-{
-	typedef std::map<std::string, void *> InstancesMap;
-	static InstancesMap instances;
-
-	// std::map::insert() is just what we want here. You attempt to insert a
-	// (key, value) pair. If the specified key doesn't yet exist, it inserts
-	// the pair and returns a std::pair of (iterator, true). If the specified
-	// key DOES exist, insert() simply returns (iterator, false). One lookup
-	// handles both cases.
-	return instances.insert(InstancesMap::value_type(info.name(),
-													 InstancesMap::mapped_type()))
-		.first->second;
-}
diff --git a/indra/llcommon/llinstancetracker.h b/indra/llcommon/llinstancetracker.h
index 1eab270e3c..55187d8325 100755
--- a/indra/llcommon/llinstancetracker.h
+++ b/indra/llcommon/llinstancetracker.h
@@ -46,21 +46,7 @@
 class LL_COMMON_API LLInstanceTrackerBase
 {
 protected:
-	/// Get a process-unique void* pointer slot for the specified type_info
-	static void * & getInstances(std::type_info const & info);
 
-	/// Find or create a STATICDATA instance for the specified TRACKED class.
-	/// STATICDATA must be default-constructible.
-	template<typename STATICDATA, class TRACKED>
-	static STATICDATA& getStatic()
-	{
-		void *& instances = getInstances(typeid(TRACKED));
-		if (! instances)
-		{
-			instances = new STATICDATA;
-		}
-		return *static_cast<STATICDATA*>(instances);
-	}
 
     /// It's not essential to derive your STATICDATA (for use with
     /// getStatic()) from StaticBase; it's just that both known
@@ -74,6 +60,8 @@ protected:
     };
 };
 
+LL_COMMON_API void assert_main_thread();
+
 /// This mix-in class adds support for tracking all instances of the specified class parameter T
 /// The (optional) key associates a value of type KEY with a given instance of T, for quick lookup
 /// If KEY is not provided, then instances are stored in a simple set
@@ -81,14 +69,18 @@ protected:
 template<typename T, typename KEY = T*>
 class LLInstanceTracker : public LLInstanceTrackerBase
 {
-	typedef LLInstanceTracker<T, KEY> MyT;
+	typedef LLInstanceTracker<T, KEY> self_t;
 	typedef typename std::map<KEY, T*> InstanceMap;
 	struct StaticData: public StaticBase
 	{
 		InstanceMap sMap;
 	};
-	static StaticData& getStatic() { return LLInstanceTrackerBase::getStatic<StaticData, MyT>(); }
-	static InstanceMap& getMap_() { return getStatic().sMap; }
+	static StaticData& getStatic() { static StaticData sData; return sData;}
+	static InstanceMap& getMap_() 
+	{
+		// assert_main_thread();   fwiw this class is not thread safe, and it used by multiple threads.  Bad things happen.
+		return getStatic().sMap; 
+	}
 
 public:
 	class instance_iter : public boost::iterator_facade<instance_iter, T, boost::forward_traversal_tag>
@@ -220,7 +212,11 @@ private:
 	}
 	void remove_()
 	{
-		getMap_().erase(mInstanceKey);
+		typename InstanceMap::iterator iter = getMap_().find(mInstanceKey);
+		if (iter != getMap_().end())
+		{
+			getMap_().erase(iter);
+		}
 	}
 
 private:
@@ -232,13 +228,13 @@ private:
 template<typename T>
 class LLInstanceTracker<T, T*> : public LLInstanceTrackerBase
 {
-	typedef LLInstanceTracker<T, T*> MyT;
+	typedef LLInstanceTracker<T, T*> self_t;
 	typedef typename std::set<T*> InstanceSet;
 	struct StaticData: public StaticBase
 	{
 		InstanceSet sSet;
 	};
-	static StaticData& getStatic() { return LLInstanceTrackerBase::getStatic<StaticData, MyT>(); }
+	static StaticData& getStatic() { static StaticData sData; return sData; }
 	static InstanceSet& getSet_() { return getStatic().sSet; }
 
 public:
diff --git a/indra/llcommon/llleap.h b/indra/llcommon/llleap.h
index 1a1ad23d39..e33f25e530 100755
--- a/indra/llcommon/llleap.h
+++ b/indra/llcommon/llleap.h
@@ -32,6 +32,7 @@
 class LL_COMMON_API LLLeap: public LLInstanceTracker<LLLeap>
 {
 public:
+
     /**
      * Pass a brief string description, mostly for logging purposes. The desc
      * need not be unique, but obviously the clearer we can make it, the
diff --git a/indra/llcommon/lllslconstants.h b/indra/llcommon/lllslconstants.h
index 9f32598e61..926ce32d75 100755
--- a/indra/llcommon/lllslconstants.h
+++ b/indra/llcommon/lllslconstants.h
@@ -67,6 +67,19 @@ const S32 LSL_PRIM_TEXGEN		= 22;
 const S32 LSL_PRIM_POINT_LIGHT	= 23;
 const S32 LSL_PRIM_CAST_SHADOWS	= 24;
 const S32 LSL_PRIM_GLOW     	= 25;
+const S32 LSL_PRIM_TEXT			= 26;
+const S32 LSL_PRIM_NAME			= 27;
+const S32 LSL_PRIM_DESC			= 28;
+const S32 LSL_PRIM_ROT_LOCAL	= 29;
+const S32 LSL_PRIM_PHYSICS_SHAPE_TYPE	= 30;
+const S32 LSL_PRIM_OMEGA		= 32;
+const S32 LSL_PRIM_POS_LOCAL	= 33;
+const S32 LSL_PRIM_LINK_TARGET	= 34;
+const S32 LSL_PRIM_SLICE		= 35;
+
+const S32 LSL_PRIM_PHYSICS_SHAPE_PRIM	= 0;
+const S32 LSL_PRIM_PHYSICS_SHAPE_NONE	= 1;
+const S32 LSL_PRIM_PHYSICS_SHAPE_CONVEX	= 2;
 
 const S32 LSL_PRIM_TYPE_BOX		= 0;
 const S32 LSL_PRIM_TYPE_CYLINDER= 1;
@@ -179,6 +192,22 @@ const S32 OBJECT_VELOCITY = 5;
 const S32 OBJECT_OWNER = 6;
 const S32 OBJECT_GROUP = 7;
 const S32 OBJECT_CREATOR = 8;
+const S32 OBJECT_RUNNING_SCRIPT_COUNT = 9;
+const S32 OBJECT_TOTAL_SCRIPT_COUNT = 10;
+const S32 OBJECT_SCRIPT_MEMORY = 11;
+const S32 OBJECT_SCRIPT_TIME = 12;
+const S32 OBJECT_PRIM_EQUIVALENCE = 13;
+const S32 OBJECT_SERVER_COST = 14;
+const S32 OBJECT_STREAMING_COST = 15;
+const S32 OBJECT_PHYSICS_COST = 16;
+const S32 OBJECT_CHARACTER_TIME = 17;
+const S32 OBJECT_ROOT = 18;
+const S32 OBJECT_ATTACHED_POINT = 19;
+const S32 OBJECT_PATHFINDING_TYPE = 20;
+const S32 OBJECT_PHYSICS = 21;
+const S32 OBJECT_PHANTOM = 22;
+const S32 OBJECT_TEMP_ON_REZ = 23;
+const S32 OBJECT_RENDER_WEIGHT = 24;
 
 // llTextBox() magic token string - yes this is a hack.  sue me.
 char const* const TEXTBOX_MAGIC_TOKEN = "!!llTextBox!!";
diff --git a/indra/llcommon/llmemory.h b/indra/llcommon/llmemory.h
index e725bdd9fa..61e30f11cc 100755
--- a/indra/llcommon/llmemory.h
+++ b/indra/llcommon/llmemory.h
@@ -36,19 +36,68 @@ class LLMutex ;
 #define LL_CHECK_MEMORY
 #endif
 
+LL_COMMON_API void ll_assert_aligned_func(uintptr_t ptr,U32 alignment);
+
+#ifdef SHOW_ASSERT
+#define ll_assert_aligned(ptr,alignment) ll_assert_aligned_func(reinterpret_cast<uintptr_t>(ptr),((U32)alignment))
+#else
+#define ll_assert_aligned(ptr,alignment)
+#endif
+
+#include <xmmintrin.h>
+
+template <typename T> T* LL_NEXT_ALIGNED_ADDRESS(T* address) 
+{ 
+	return reinterpret_cast<T*>(
+		(reinterpret_cast<uintptr_t>(address) + 0xF) & ~0xF);
+}
+
+template <typename T> T* LL_NEXT_ALIGNED_ADDRESS_64(T* address) 
+{ 
+	return reinterpret_cast<T*>(
+		(reinterpret_cast<uintptr_t>(address) + 0x3F) & ~0x3F);
+}
+
+#if LL_LINUX || LL_DARWIN
+
+#define			LL_ALIGN_PREFIX(x)
+#define			LL_ALIGN_POSTFIX(x)		__attribute__((aligned(x)))
+
+#elif LL_WINDOWS
+
+#define			LL_ALIGN_PREFIX(x)		__declspec(align(x))
+#define			LL_ALIGN_POSTFIX(x)
+
+#else
+#error "LL_ALIGN_PREFIX and LL_ALIGN_POSTFIX undefined"
+#endif
+
+#define LL_ALIGN_16(var) LL_ALIGN_PREFIX(16) var LL_ALIGN_POSTFIX(16)
+
 inline void* ll_aligned_malloc( size_t size, int align )
 {
+#if defined(LL_WINDOWS)
+	return _aligned_malloc(size, align);
+#else
 	void* mem = malloc( size + (align - 1) + sizeof(void*) );
 	char* aligned = ((char*)mem) + sizeof(void*);
 	aligned += align - ((uintptr_t)aligned & (align - 1));
 
 	((void**)aligned)[-1] = mem;
 	return aligned;
+#endif
 }
 
 inline void ll_aligned_free( void* ptr )
 {
-	free( ((void**)ptr)[-1] );
+#if defined(LL_WINDOWS)
+	_aligned_free(ptr);
+#else
+	if (ptr)
+	{
+		free( ((void**)ptr)[-1] );
+	}
+#endif
 }
 
 #if !LL_USE_TCMALLOC
@@ -133,6 +182,78 @@ inline void ll_aligned_free_32(void *p)
 #endif
 }
 
+
+// Copy words 16-byte blocks from src to dst. Source and destination MUST NOT OVERLAP. 
+// Source and dest must be 16-byte aligned and size must be multiple of 16.
+//
+inline void ll_memcpy_nonaliased_aligned_16(char* __restrict dst, const char* __restrict src, size_t bytes)
+{
+	assert(src != NULL);
+	assert(dst != NULL);
+	assert(bytes > 0);
+	assert((bytes % sizeof(F32))== 0); 
+	ll_assert_aligned(src,16);
+	ll_assert_aligned(dst,16);
+	assert((src < dst) ? ((src + bytes) < dst) : ((dst + bytes) < src));
+	assert(bytes%16==0);
+
+	char* end = dst + bytes;
+
+	if (bytes > 64)
+	{
+
+		// Find start of 64b aligned area within block
+		//
+		void* begin_64 = LL_NEXT_ALIGNED_ADDRESS_64(dst);
+		
+		//at least 64 bytes before the end of the destination, switch to 16 byte copies
+		void* end_64 = end-64;
+	
+		// Prefetch the head of the 64b area now
+		//
+		_mm_prefetch((char*)begin_64, _MM_HINT_NTA);
+		_mm_prefetch((char*)begin_64 + 64, _MM_HINT_NTA);
+		_mm_prefetch((char*)begin_64 + 128, _MM_HINT_NTA);
+		_mm_prefetch((char*)begin_64 + 192, _MM_HINT_NTA);
+	
+		// Copy 16b chunks until we're 64b aligned
+		//
+		while (dst < begin_64)
+		{
+
+			_mm_store_ps((F32*)dst, _mm_load_ps((F32*)src));
+			dst += 16;
+			src += 16;
+		}
+	
+		// Copy 64b chunks up to your tail
+		//
+		// might be good to shmoo the 512b prefetch offset
+		// (characterize performance for various values)
+		//
+		while (dst < end_64)
+		{
+			_mm_prefetch((char*)src + 512, _MM_HINT_NTA);
+			_mm_prefetch((char*)dst + 512, _MM_HINT_NTA);
+			_mm_store_ps((F32*)dst, _mm_load_ps((F32*)src));
+			_mm_store_ps((F32*)(dst + 16), _mm_load_ps((F32*)(src + 16)));
+			_mm_store_ps((F32*)(dst + 32), _mm_load_ps((F32*)(src + 32)));
+			_mm_store_ps((F32*)(dst + 48), _mm_load_ps((F32*)(src + 48)));
+			dst += 64;
+			src += 64;
+		}
+	}
+
+	// Copy remainder 16b tail chunks (or ALL 16b chunks for sub-64b copies)
+	//
+	while (dst < end)
+	{
+		_mm_store_ps((F32*)dst, _mm_load_ps((F32*)src));
+		dst += 16;
+		src += 16;
+	}
+}
+
 #ifndef __DEBUG_PRIVATE_MEM__
 #define __DEBUG_PRIVATE_MEM__  0
 #endif
@@ -541,13 +662,7 @@ void  LLPrivateMemoryPoolTester::operator delete[](void* addr)
 
 // LLSingleton moved to llsingleton.h
 
-LL_COMMON_API void ll_assert_aligned_func(uintptr_t ptr,U32 alignment);
 
-#ifdef SHOW_ASSERT
-#define ll_assert_aligned(ptr,alignment) ll_assert_aligned_func(reinterpret_cast<uintptr_t>(ptr),((U32)alignment))
-#else
-#define ll_assert_aligned(ptr,alignment)
-#endif
 
 
 #endif
diff --git a/indra/llcommon/llsingleton.cpp b/indra/llcommon/llsingleton.cpp
index eb8e2c9456..9b49e52377 100755
--- a/indra/llcommon/llsingleton.cpp
+++ b/indra/llcommon/llsingleton.cpp
@@ -28,5 +28,4 @@
 
 #include "llsingleton.h"
 
-std::map<std::string, void *> * LLSingletonRegistry::sSingletonMap = NULL;
 
diff --git a/indra/llcommon/llsingleton.h b/indra/llcommon/llsingleton.h
index 49d99f2cd0..40002313f1 100755
--- a/indra/llcommon/llsingleton.h
+++ b/indra/llcommon/llsingleton.h
@@ -30,38 +30,6 @@
 #include <typeinfo>
 #include <boost/noncopyable.hpp>
 
-/// @brief A global registry of all singletons to prevent duplicate allocations
-/// across shared library boundaries
-class LL_COMMON_API LLSingletonRegistry {
-	private:
-		typedef std::map<std::string, void *> TypeMap;
-		static TypeMap * sSingletonMap;
-
-		static void checkInit()
-		{
-			if(sSingletonMap == NULL)
-			{
-				sSingletonMap = new TypeMap();
-			}
-		}
-
-	public:
-		template<typename T> static void * & get()
-		{
-			std::string name(typeid(T).name());
-
-			checkInit();
-
-			// the first entry of the pair returned by insert will be either the existing
-			// iterator matching our key, or the newly inserted NULL initialized entry
-			// see "Insert element" in http://www.sgi.com/tech/stl/UniqueAssociativeContainer.html
-			TypeMap::iterator result =
-				sSingletonMap->insert(std::make_pair(name, (void*)NULL)).first;
-
-			return result->second;
-		}
-};
-
 // LLSingleton implements the getInstance() method part of the Singleton
 // pattern. It can't make the derived class constructors protected, though, so
 // you have to do that yourself.
@@ -101,20 +69,23 @@ private:
 	} EInitState;
 	
 	// stores pointer to singleton instance
-	// and tracks initialization state of singleton
-	struct SingletonInstanceData
+	struct SingletonLifetimeManager
 	{
-		EInitState		mInitState;
-		DERIVED_TYPE*	mSingletonInstance;
-		
-		SingletonInstanceData()
-		:	mSingletonInstance(NULL),
-			mInitState(UNINITIALIZED)
-		{}
-
-		~SingletonInstanceData()
+		SingletonLifetimeManager()
+		{
+			construct();
+		}
+
+		static void construct()
 		{
-			if (mInitState != DELETED)
+			sData.mInitState = CONSTRUCTING;
+			sData.mInstance = new DERIVED_TYPE(); 
+			sData.mInitState = INITIALIZING;
+		}
+
+		~SingletonLifetimeManager()
+		{
+			if (sData.mInitState != DELETED)
 			{
 				deleteSingleton();
 			}
@@ -124,9 +95,8 @@ private:
 public:
 	virtual ~LLSingleton()
 	{
-		SingletonInstanceData& data = getData();
-		data.mSingletonInstance = NULL;
-		data.mInitState = DELETED;
+		sData.mInstance = NULL;
+		sData.mInitState = DELETED;
 	}
 
 	/**
@@ -151,50 +121,46 @@ public:
 	 */
 	static void deleteSingleton()
 	{
-		delete getData().mSingletonInstance;
-		getData().mSingletonInstance = NULL;
-		getData().mInitState = DELETED;
+		delete sData.mInstance;
+		sData.mInstance = NULL;
+		sData.mInitState = DELETED;
 	}
 
-	static SingletonInstanceData& getData()
-	{
-		// this is static to cache the lookup results
-		static void * & registry = LLSingletonRegistry::get<DERIVED_TYPE>();
-
-		// *TODO - look into making this threadsafe
-		if(NULL == registry)
-		{
-			static SingletonInstanceData data;
-			registry = &data;
-		}
-
-		return *static_cast<SingletonInstanceData *>(registry);
-	}
 
 	static DERIVED_TYPE* getInstance()
 	{
-		SingletonInstanceData& data = getData();
+		static SingletonLifetimeManager sLifeTimeMgr;
 
-		if (data.mInitState == CONSTRUCTING)
+		switch (sData.mInitState)
 		{
+		case UNINITIALIZED:
+			// should never be uninitialized at this point
+			llassert(false);
+			return NULL;
+		case CONSTRUCTING:
 			llerrs << "Tried to access singleton " << typeid(DERIVED_TYPE).name() << " from singleton constructor!" << llendl;
-		}
-
-		if (data.mInitState == DELETED)
-		{
+			return NULL;
+		case INITIALIZING:
+			// go ahead and flag ourselves as initialized so we can be reentrant during initialization
+			sData.mInitState = INITIALIZED;	
+			sData.mInstance->initSingleton(); 
+			return sData.mInstance;
+		case INITIALIZED:
+			return sData.mInstance;
+		case DELETED:
 			llwarns << "Trying to access deleted singleton " << typeid(DERIVED_TYPE).name() << " creating new instance" << llendl;
+			SingletonLifetimeManager::construct();
+			sData.mInitState = INITIALIZED;	
+			sData.mInstance->initSingleton(); 
+			return sData.mInstance;
 		}
-		
-		if (!data.mSingletonInstance) 
-		{
-			data.mInitState = CONSTRUCTING;
-			data.mSingletonInstance = new DERIVED_TYPE(); 
-			data.mInitState = INITIALIZING;
-			data.mSingletonInstance->initSingleton(); 
-			data.mInitState = INITIALIZED;	
-		}
-		
-		return data.mSingletonInstance;
+
+		return NULL;
+	}
+
+	static DERIVED_TYPE* getIfExists()
+	{
+		return sData.mInstance;
 	}
 
 	// Reference version of getInstance()
@@ -208,18 +174,29 @@ public:
 	// Use this to avoid accessing singletons before the can safely be constructed
 	static bool instanceExists()
 	{
-		return getData().mInitState == INITIALIZED;
+		return sData.mInitState == INITIALIZED;
 	}
 	
 	// Has this singleton already been deleted?
 	// Use this to avoid accessing singletons from a static object's destructor
 	static bool destroyed()
 	{
-		return getData().mInitState == DELETED;
+		return sData.mInitState == DELETED;
 	}
 
 private:
+
 	virtual void initSingleton() {}
+
+	struct SingletonData
+	{
+		EInitState		mInitState;
+		DERIVED_TYPE*	mInstance;
+	};
+	static SingletonData sData;
 };
 
+template<typename T>
+typename LLSingleton<T>::SingletonData LLSingleton<T>::sData;
+
 #endif
diff --git a/indra/llcommon/llstaticstringtable.h b/indra/llcommon/llstaticstringtable.h
new file mode 100644
index 0000000000..d7e0e8a08d
--- /dev/null
+++ b/indra/llcommon/llstaticstringtable.h
@@ -0,0 +1,82 @@
+/** 
+ * @file llstringtable.h
+ * @brief The LLStringTable class provides a _fast_ method for finding
+ * unique copies of strings.
+ *
+ * $LicenseInfo:firstyear=2001&license=viewerlgpl$
+ * Second Life Viewer Source Code
+ * Copyright (C) 2010, Linden Research, Inc.
+ * 
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation;
+ * version 2.1 of the License only.
+ * 
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ * 
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this library; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
+ * 
+ * Linden Research, Inc., 945 Battery Street, San Francisco, CA  94111  USA
+ * $/LicenseInfo$
+ */
+
+#ifndef LL_STATIC_STRING_TABLE_H
+#define LL_STATIC_STRING_TABLE_H
+
+#include "lldefs.h"
+#include <boost/unordered_map.hpp>
+#include "llstl.h"
+
+class LLStaticHashedString
+{
+public:
+
+	LLStaticHashedString(const std::string& s)
+	{
+		string_hash = makehash(s);
+		string		= s;
+	}
+
+	const std::string&	String() const { return string;		}
+	size_t				Hash()	 const { return string_hash;  }
+
+	bool operator==(const LLStaticHashedString& b) const { return Hash() == b.Hash(); }
+
+protected:
+
+	size_t makehash(const std::string& s)
+	{
+		size_t len = s.size();
+		const char* c = s.c_str();
+		size_t hashval = 0;
+		for (size_t i=0; i<len; i++)
+		{
+			hashval = ((hashval<<5) + hashval) + *c++;
+		}
+		return hashval;
+	}
+
+	std::string string;
+	size_t		string_hash;
+};
+
+struct LLStaticStringHasher
+{
+	enum { bucket_size = 8 };
+	size_t operator()(const LLStaticHashedString& key_value) const { return key_value.Hash(); }
+	bool   operator()(const LLStaticHashedString& left, const LLStaticHashedString& right) const { return left.Hash() < right.Hash(); }
+};
+
+template< typename MappedObject >
+class LL_COMMON_API LLStaticStringTable
+	: public boost::unordered_map< LLStaticHashedString, MappedObject, LLStaticStringHasher >
+{
+};
+
+#endif
+
diff --git a/indra/llcommon/llstringtable.h b/indra/llcommon/llstringtable.h
index 59d7372ed4..ff09e71677 100755
--- a/indra/llcommon/llstringtable.h
+++ b/indra/llcommon/llstringtable.h
@@ -42,14 +42,6 @@
 //# define STRING_TABLE_HASH_MAP 1
 #endif
 
-#if STRING_TABLE_HASH_MAP
-# if LL_WINDOWS
-#  include <hash_map>
-# else
-#  include <ext/hash_map>
-# endif
-#endif
-
 const U32 MAX_STRINGS_LENGTH = 256;
 
 class LL_COMMON_API LLStringTableEntry
diff --git a/indra/llcommon/llthread.cpp b/indra/llcommon/llthread.cpp
index 1d56a52c32..60adeeaeb7 100755
--- a/indra/llcommon/llthread.cpp
+++ b/indra/llcommon/llthread.cpp
@@ -67,7 +67,8 @@ LL_COMMON_API void assert_main_thread()
 	static U32 s_thread_id = LLThread::currentID();
 	if (LLThread::currentID() != s_thread_id)
 	{
-		llerrs << "Illegal execution outside main thread." << llendl;
+		llwarns << "Illegal execution from thread id " << (S32) LLThread::currentID()
+			<< " outside main thread " << (S32) s_thread_id << llendl;
 	}
 }
 
@@ -495,15 +496,7 @@ LLThreadSafeRefCount::LLThreadSafeRefCount() :
 
 LLThreadSafeRefCount::LLThreadSafeRefCount(const LLThreadSafeRefCount& src)
 {
-	if (sMutex)
-	{
-		sMutex->lock();
-	}
 	mRef = 0;
-	if (sMutex)
-	{
-		sMutex->unlock();
-	}
 }
 
 LLThreadSafeRefCount::~LLThreadSafeRefCount()
diff --git a/indra/llcommon/llthread.h b/indra/llcommon/llthread.h
index 0fb89c5613..f51d985b5f 100755
--- a/indra/llcommon/llthread.h
+++ b/indra/llcommon/llthread.h
@@ -242,49 +242,39 @@ public:
 	LLThreadSafeRefCount(const LLThreadSafeRefCount&);
 	LLThreadSafeRefCount& operator=(const LLThreadSafeRefCount& ref) 
 	{
-		if (sMutex)
-		{
-			sMutex->lock();
-		}
 		mRef = 0;
-		if (sMutex)
-		{
-			sMutex->unlock();
-		}
 		return *this;
 	}
 
-
-	
 	void ref()
 	{
-		if (sMutex) sMutex->lock();
 		mRef++; 
-		if (sMutex) sMutex->unlock();
 	} 
 
-	S32 unref()
+	void unref()
 	{
 		llassert(mRef >= 1);
-		if (sMutex) sMutex->lock();
-		S32 res = --mRef;
-		if (sMutex) sMutex->unlock();
-		if (0 == res) 
-		{
-			delete this; 
-			return 0;
+		if ((--mRef) == 0)		// See note in llapr.h on atomic decrement operator return value.  
+		{	
+			// If we hit zero, the caller should be the only smart pointer owning the object and we can delete it.
+			// It is technically possible for a vanilla pointer to mess this up, or another thread to
+			// jump in, find this object, create another smart pointer and end up dangling, but if
+			// the code is that bad and not thread-safe, it's trouble already.
+			delete this;
 		}
-		return res;
-	}	
+	}
+
 	S32 getNumRefs() const
 	{
-		return mRef;
+		const S32 currentVal = mRef.CurrentValue();
+		return currentVal;
 	}
 
 private: 
-	S32	mRef; 
+	LLAtomic32< S32	> mRef; 
 };
 
+
 /**
  * intrusive pointer support for LLThreadSafeRefCount
  * this allows you to use boost::intrusive_ptr with any LLThreadSafeRefCount-derived type
@@ -315,4 +305,6 @@ public:
 
 //============================================================================
 
+extern LL_COMMON_API void assert_main_thread();
+
 #endif // LL_LLTHREAD_H